Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Nov 4.
Published in final edited form as: Drug Alcohol Depend. 2026 Oct 16;288:113400. doi: 10.1016/j.drugalcdep.2026.113400

Neighborhood Violence, Visible Drug Activity, and Substance Use among Urban Youth: Mental Health and Community Factors

Sitara M Weerakoon 1, Ijeoma Opara 1
PMCID: PMC13729271  NIHMSID: NIHMS2218958  PMID: 42913353

Abstract

Background

Adolescent substance use and exposure to neighborhood violence frequently co-occur in U.S. urban settings, particularly among racially minoritized and low-income youth. Despite shared exposure to environmental risk, some adolescents report low to no substance use. Guided by the adverse community experiences and resilience framework, this study identified classes of neighborhood risk and substance use among urban adolescents and examined whether emotional, behavioral, community, and demographic factors differentiated profile membership.

Methods

Cross-sectional survey data were collected from 743 adolescents aged 13–21 residing in a large Northeastern U.S. city. Five dichotomous indicators (witnessed neighborhood violence, seen drugs sold in their neighborhood, and lifetime vaping, alcohol, and marijuana use) were analyzed using latent class analysis. Multinomial logistic regression examined associations between class membership and emotional, behavioral, community, and demographic factors.

Results

A three-class solution was retained: No Neighborhood Risk Exposure/Low or No Substance Use (34%), High Exposure/Low or No Use (40%), and High Exposure/Polysubstance Use (26%). Compared with the High Exposure/Low or No Use class, the No Neighborhood Risk Exposure/Low or No Use class reported higher neighborhood support and lower racial discrimination, missing school, and suicidal thoughts. The High Exposure/Polysubstance Use class differed from the High Exposure/Low or No Use class primarily by higher depressive symptoms.

Conclusions

Low substance use should not be interpreted as low need among adolescents exposed to neighborhood risk. Adolescent health and well-being efforts should assess mental health distress, suicidal ideation, racial discrimination-related experiences, and school disengagement alongside substance use.

Keywords: neighborhood violence, adolescents, drug and alcohol use, polysubstance use, adolescent health

1. Introduction

1.1. Background

Adolescent substance use and exposure to community violence remain significant public health concerns in urban areas of the United States (US). National data indicates that nearly one in three US high school students report current alcohol use, one in five report past-month marijuana use, and more than 10% report current electronic vapor product use ((CDC), 2025). While each neighborhood is unique and have different rates, current data have found that 23% of youth nationally have ever seen someone get physically attacked, beaten, stabbed, or shot in their neighborhood ((CDC), 2025; Hart et al., 2025). Prior data has estimated that 60% of youth in urban communities report witnessing or directly experiencing some form of violence during adolescence (Finkelhor et al., 2015; Zimmerman & Messner, 2013). These exposures disproportionately burden racially minoritized and low-income populations due to structural inequities such as historical residential segregation, housing policies, and resource allocation (Alaniz, 1998). These structural inequities also influence environmental risk factors such as visible drug activity and violent crime (Graif et al., 2014; Lardier et al., 2021). Exposure to and normalization of behaviors and conditions like drug activity and crime can influence initiation of substance use behaviors in youth and young adulthood (Guy et al., 1994; Motyka & Al-Imam, 2022).

Adolescents exposed to these environments frequently navigate multiple layers of challenges and supports that shape their behavioral and emotional responses. Racial discrimination is an example of a risk factor, and neighborhood/community support (i.e., mutual trust and cohesion within a community) is an example of a protective factor which play an important role in how young people adapt to neighborhood stressors (Kingsbury et al., 2020; Wang et al., 2020). Experiences of racial discrimination for example can contribute to chronic stress and emotional dysregulation, which have been linked to substance use as a coping strategy among racially minoritized youth (Heads et al., 2021; Qeadan et al., 2022). Higher levels of neighborhood support are associated with lower rates of youth violence exposure and low substance use, serving as a protective community-level factor (Danielsson, 2021). Together, these factors shape how adolescents internalize and respond to neighborhood risk through both behavioral and emotional pathways.

Exposure to community violence and visible drug activity is associated with a range of negative behaviors and conditions, including substance use, anxiety, depression, and suicidality (Lambert et al., 2008; Opara I et al., 2025; Opara et al., 2020; Opara et al., 2021; Vermeiren et al., 2002; Weerakoon et al., 2023). Qualitative evidence suggests that the constant surrounding of environmental conditions, including drug activity and violence, can normalize substance use for adolescents (Opara et al., 2026). However, not all adolescents exposed to these conditions engage in substance use, suggesting the presence of resilience processes or adaptive coping mechanisms (Anona et al., 2024; Perreira et al., 2019). Delineating divergence in these experiences, as well as understanding the risk and protective factors for this divergence, is critical to identifying both risk and resilience pathways within high-risk urban contexts.

1.2. Theoretical framework

The adverse community experiences and resilience (ACER) framework provides a useful lens for examining how neighborhood conditions shape adolescent substance use, mental health, and community experiences. The ACER framework extends trauma theory beyond the individual level by positing that trauma can manifest at the community or population level, particularly in communities affected by interpersonal violence, structural violence, disinvestment, and weakened social networks (Pinderhughes et al., 2015). Rather than conceptualizing trauma solely as an individual clinical condition, ACER emphasizes that community trauma is reflected in the 1) social-cultural, 2) physical, and 3) economic environments. The social-cultural environment involves social processes that concentrate poverty in inner city neighborhoods which damage things like trust in systems and the ability to take action for change. The physical environment involves physical decay and decline which are an added element of toxic stress to residents. The economic environments involve the process through which neighborhood characteristics he concentration of poverty levels of violence, crime and delinquency, education, psychological distress, and various health problems are affected by neighborhood characteristics, particularly the concentration of poverty. Importantly, the ACER framework recognizes that community trauma may be experienced and expressed differently across residents, supporting attention to heterogeneity in youth experiences. This framework supports the present study in two ways. First, the ACER framework provides a conceptual lens for interpreting youth-reported exposure to characteristics of neighborhood violence and disorder in relation to their well-being, without assuming a uniform response across youth (Pinderhughes et al., 2015). Second, the ACER framework emphasizes resilience-promoting features of communities, including trust and mutual support. Together, these principles justify a person-centered approach to identifying patterns of neighborhood exposure and substance use and examining whether these patterns are characterized by different behavioral, emotional, and community-related experiences.

1.3. Present Study

Not all adolescents living within environments burdened with drug activity and violence engage in substance use. Therefore, to delineate and understand this divergence the present study sought to identify cross-sectional classes of conditions of neighborhood violence, visible drug activity, and substance use behaviors among urban adolescents, and to characterize these classes by behavioral, emotional, community, and demographic factors. Two primary aims guided this analysis:

  1. To identify latent classes of neighborhood conditions and substance use patterns based on observed indicators of violence exposure, visible drug activity, and lifetime use of vaping, alcohol, and marijuana.

  2. To examine whether experiences of neighborhood support, racial discrimination, school engagement, mental health indicators (depressive symptoms, anxiety, and suicidality), and demographics differ across these classes.

By distinguishing subgroups of adolescents with unique patterns of environmental and behavioral experiences, this study seeks to clarify mechanisms of adaptation in the face of neighborhood-level trauma. Understanding this can clarify how adolescents in high-risk environments diverge in their behavioral and emotional responses.

2. Methods

2.1. Setting and Participants

Data for this cross-sectional analysis were drawn from a community-based youth participatory action study conducted in a large urban city in the Northeastern US. The study was focused on developing a substance use intervention by youth for youth. A youth advisory board (YAB) informed data collection, analysis, and interpretation of results. The city is ethnically diverse, with a population that is approximately 60% Hispanic/Latinx and 25% Black/African American, and experiences disproportionate rates of poverty and neighborhood violence relative to the state average(US Census Bureau - QuickFacts Paterson, NJ, 2021). Adolescents aged 13-21 years were recruited through partnerships with community-based organizations, schools, and youth-serving events using venue-based, purposive, and snowball sampling methods. These methods are appropriate community-based studies for which reaching students who may be absent or truant from school is a priority (Biernacki & Waldorf, 1981; Etikan et al., 2016; Opara et al., 2021; Verdery et al., 2019). Eligible participants were required to (1) reside in the study city, (2) be within the eligible age range, and (3) provide assent to participate. Parental consent was waived in accordance with state policy for minimal-risk research on sensitive health topics to ensure confidentiality and accurate reporting. Youth completed electronic surveys using Qualtrics tablets or iPads in community settings. Participants received a $10 incentive for completion. The institutional review board approved all procedures. A total of 769 youth were initially recruited, and after excluding those who did not qualify, 743 participants completed the survey which represented the final sample size.

2.2. Measures

Consistent with ACER framework and informed by community-engaged discussions with youth where some described abstaining from substance use despite witnessing drug activity and violence, five dichotomous indicators were used to identify latent classes of neighborhood risk exposure and substance use to capture the co-occurrence of environmental and behavioral conditions. Participants were asked whether they had ever seen drugs sold in their neighborhood (“Have you ever seen someone sell drugs in your neighborhood?”; 0 = No, 1 = Yes) and whether they had ever witnessed neighborhood violence (“Have you ever seen someone get attacked, beaten, stabbed, or shot in your neighborhood?”; 0 = No, 1 = Yes). Lifetime drug (vaping, alcohol, and marijuana) use were each assessed using standard Youth Risk Behavior Surveillance System (YRBSS) (Brener, 2024) items asking, “Have you ever used an electronic vapor product (such as JUUL, Ruyan, or NJOY)?” (0 = Never, 1 = Yes), “Have you ever used marijuana? (0 = Never, 1 = Yes), and “During your life, how many times have you used [alcohol]? (0 = Never, 1 = At least once). Here, substance use represents a behavioral domain that was modeled with neighborhood exposures to identify groups of youth who reported similar environmental conditions but may have differed in substance-use initiation. This distinction allows us to examine whether behavioral, emotional, and community-related experiences differentiate youth with varying configurations of neighborhood exposure and substance use.

To examine factors associated with class membership, several community, behavioral, emotional, and demographic factors were included.

Community factors

Community factors captured youth perceptions of and experiences within their broader social and neighborhood environments. Neighborhood willingness to help (hereafter referred to as neighborhood support for brevity) was measured by agreement with the statement, “People in my neighborhood are willing to help their neighbors,” on a five-point Likert scale (1 = strongly disagree to 5 = strongly agree), treated as a continuous variable. Perceived availability of community youth activities was measured with the question, “Which of the following activities for people your age are available in your community?” with responses options including 1) sports teams, 2) scouting (Boy Scouts or Girl Scouts), 3) Boys and Girls Clubs, 4) religious groups or church youth groups, 5) service clubs, community service groups, or other groups aimed at helping others, and 6) organized clubs such as band, choir, drill team or drama club. Participants could select as many as were relevant; items were summed to a total of six possible activities and measured on a continuous scale. Racial discrimination was measured with the question, “In the past year (12 months), how much negative discrimination have you experienced because of your race?” (0 = None, 1 = At least some).

Behavioral factor

Missing school was measured by the question, “During the last four weeks, how many whole days of school have you missed?” and dichotomized as 0 = no missed days, and 1 = missed ≥1 day. Missing school was categorized as an behavioral factor because school absenteeism is commonly treated as a behavioral indicator of school disengagement and has been linked to behavioral problems such as behavioral dysregulation and conduct-related risk (Hendron & Kearney, 2016; Olivier et al., 2020). However, absenteeism may also reflect emotional challenges, family responsibilities, transportation barriers, school safety concerns, or other structural constraints (Burdick-Will et al., 2019; Fornander & Kearney, 2019; Lim et al., 2019); therefore, this measure is interpreted as a broad behavioral marker of school disengagement.

Emotional factors

Mental health was evaluated using the Patient Health Questionnaire (PHQ-9) (Fonseca-Pedrero et al., 2023) and the Generalized Anxiety Disorder Scale (GAD-7) (Löwe et al., 2008), both treated as continuous measures of depressive and anxiety symptoms, respectively, with the PHQ-9 total score ranging from 0-27 and the GAD-7 total score ranging from 0-21. Suicidal thoughts was assessed with the question, “Has there been a time in the past month when you have had serious thoughts about ending your life?” (0=No, 1=Yes). Demographic variables included age (in years), gender (0 = male, 1 = female), and race/ethnicity (Black, Latinx, Multiracial, or Other).

2.3. Analytic Procedures

Descriptive analyses were conducted to summarize sample characteristics. Latent class analysis (LCA) was performed in Stata statistical software version 18 using the generalized structural equation modeling (GSEM) framework to identify unobserved subgroups based on the five observed indicators. Model selection was based on statistical fit (AIC, BIC, entropy), interpretability, and theoretical relevance. Three classes were retained: a low-neighborhood exposure/low substance use class, a high-exposure/low-use class, and a high-exposure/high-use class. Participants were assigned to their most probable latent class based on posterior probabilities. To assess differences in community, behavioral, emotional, and demographic factors by latent class membership, a contrast approach was taken in which multinomial logistic regression models were estimated with alternating referent categories. Models were first estimated with Class 1 as the referent group and then re-estimated with Class 2 as the referent group, allowing pairwise comparisons across classes. Correlates included support from neighbors, perceived availability of community youth activities, suicidal thoughts, missing school, depression symptoms, anxiety symptoms, and demographics. All analyses were conducted with Stata statistical software version 18.5 (College Station, TX). Statistical significance was set at alpha=0.05.

Missing data was minimal across variables in the data set (<5%). Stata applies a full-information maximum likelihood (FIML) when running LCA via GSEM, allowing all available data to contribute to parameter estimation without listwise deletion; the analytic sample for the GSEM models was n=726. For logistic regression, Stata applies a complete case analysis approach; the final analytic sample for the multinomial logistic regression models was n=544. To assess potential selection associated with complete-case estimation, participants included in the multinomial logistic regression analysis were compared with LCA participants excluded because of missing covariate data. Chi-square tests were used to compare categorical characteristics, and an independent-samples t-test was used to compare age. We also examined whether the proportion retained in the regression analysis differed across the three assigned latent classes.

3. Results

Table 1 summarizes demographic, psychosocial, and contextual characteristics of the analytic sample (n = 743). Participants were on average 15.5 years old (SD = 1.7), and slightly more than half identified as male (53.5%). The majority identified as Latinx (54.3%), followed by Black/African American (31.0%), Multiracial (7.4%), and other racial groups (7.3%). Approximately 15% of adolescents reported having serious suicidal thoughts in the past year. The mean scores for depression (PHQ-9 = 5.5, SD = 6.8) and anxiety (GAD-7 = 5.3, SD = 5.9) symptoms reflected mild symptomatology on average. Nearly half (45.5%) of youth reported experiencing at least one instance of racial discrimination, and the average ‘neighbors willing to help’ score was 2.7 (SD = 0.9), indicating moderate perceptions of neighborhood support. Approximately 18% reported missing at least one day of school in the past month. Lifetime substance use was reported by 22.8% of the sample for vaping, 26.0% for alcohol, and 20.0% for marijuana. Neighborhood-level exposures were highly prevalent; 43.0% had seen drugs sold and 42.3% had witnessed physical assault or violence in their neighborhoods.

Table 1.

Descriptive statistics of the adolescent study sample (n=743)

N (%) or Mean (SD)
Demographic factors
Age (years) (mean (SD)) 15.5 (1.7)
Sex
Male 378 (53.5)
Female 328 (46.5)
Race/ethnicity
Black/African American 226 (31.0)
Other 53 (7.3)
Latinx 395 (54.3)
Multiracial 54 (7.4)
Emotional factors
Suicidal thoughts (yes) 108 (15.0)
Anxiety symptoms (GAD-7) (mean (SD)) 5.3 (5.9)
Depressive symptoms (PHQ-9) (mean (SD)) 5.5 (6.8)
Behavioral factor
Missing school (at least one day)
At least one day 129 (18.3)
Zero days 575 (81.7)
Community factors
Neighborhood support (mean (SD)) 2.7 (0.9)
Perceived availability of community youth activities (median (range) 4 (6)
Racial discrimination
(≥ 1 instance) 325 (45.5)
No instances 389 (54.5)
Variables for LCA
Lifetime vape use (ever) 166 (22.8)
Lifetime alcohol use (ever) 189 (26.0)
Lifetime marijuana use (ever) 146 (20.0)
Witnessing assault and violence in the neighborhood (yes) 206 (42.3)
Witnessing drugs sold in the neighborhood (yes) 309 (43.0)
Open in a new tab

Abbreviations: SD, standard deviation, GAD-7, generalized anxiety disorder scale, PHQ-9, patient health questionnaire, LCA, latent class analysis

Notes: Availability of community youth activities is a count variable indicating how many activities are available to youth ranging from 0 to 6.

Missing data presented if missing>5%.

Model fit of the latent class analysis results are presented in Table 2. Model fit statistics supported a three-class solution as the most parsimonious and theoretically meaningful representation of the data. The three-class model demonstrated improved fit (AIC = 3845.90; BIC = 3919.30) compared with the two-class model (AIC = 3898.74; BIC = 3949.20) and marginal improvement beyond a four-class solution (AIC = 3836.05; BIC = 3923.21), while maintaining interpretability. Although the four-class model had a slightly lower AIC, the additional class did not yield a substantively distinct or theoretically interpretable subgroup; therefore, the three-class model was retained as the most parsimonious solution.

Table 2.

Latent class model fit statistics for the youth sample (n=726).

LL df AIC BIC
2-class model −1938.369 11 3898.739 3949.202
3-class model −1906.948 16 3845.895 3919.296
4-class model −1899.023 19 3836.046 3923.209
Open in a new tab

Figure 1 illustrates the conditional probabilities of each latent indicator across the three classes. Youth in the No Exposure/Low or No Substance Use class showed low probabilities (< 0.20) of both neighborhood exposure and lifetime substance use across all indicators. The High Exposure/Low or No Use class displayed high probabilities (> 0.70) of witnessing neighborhood violence and drug sales but low probabilities of lifetime substance use (< 0.30). In contrast, the High Exposure/Polysubstance Use class exhibited consistently high probabilities of both environmental exposure and substance use (> 0.70 across vaping, alcohol, and marijuana indicators). Estimated class proportions were 34% for Class 1, 40% for Class 2, and 26% for Class 3. Entropy of this selected model was 0.91 indicating high class separation.

Figure 1.

Figure 1

Open in a new tab

Latent class prohabilities of observed variables in the youth sample (n=726)

Notes: Estimated class proportions were 34% for Class 1, 40% for Class 2, and 26% for Class 3.

Multinomial logistic regression results in Table 3 identified several factors that differentiated latent class membership across targeted class contrasts. Compared with adolescents in the High Exposure/Low or No Use class, adolescents in the No Exposure/Low or No Use class reported significantly higher neighborhood support (RRR = 1.32, 95% CI = 1.04–1.68), lower likelihood of racial discrimination (RRR = 0.32, 95% CI = 0.20–0.51), lower likelihood of missing school (RRR = 0.52, 95% CI = 0.28–0.98), and lower likelihood of suicidal thoughts (RRR = 0.42, 95% CI = 0.20–0.88). Compared with the High Exposure/Low or No Use class, membership in the High Exposure/Polysubstance Use class was associated with greater depressive symptoms (RRR = 1.09, 95% CI = 1.03–1.15) and lower likelihood of identifying as ‘Other’ race (compared to Black adolescents) (RRR = 0.08, 95% CI = 0.01–0.64). Compared to the No Exposure/No or Low Use Class, adolescents in the High Exposure/Polysubstance Use class were more likely to report racial discrimination (RRR = 2.53, 95% CI = 1.49–4.32), missing school (RRR = 2.95, 95% CI = 1.51–5.76), greater depressive symptoms (RRR = 1.06, 95% CI = 1.01–1.14), and older age (RRR = 1.22, 95% CI = 1.04–1.43). Adolescents identifying as ‘Other’ race were less likely than Black adolescents to belong to the High Exposure/ Polysubstance Use class compared with the No Exposure/Low or No Use class (RRR = 0.04, 95% CI = 0.01–0.35). Regarding the reduction in sample size from the LCA to the multinomial regression analysis, complete-case retention differed across the three latent profiles and additionally by age, gender, and exposure to visible neighborhood drug activity, but not by race/ethnicity, witnessing neighborhood assault, or lifetime vaping, marijuana, or alcohol use (results not shown).

Table 3.

Multinomial logistic regression comparing latent classes using targeted contrasts (n=544).

Panel A
Class 1 vs Class 2
(No Exposure/Low or No Use vs High Exposure/Low or No Use)		 Panel B
Class 3 vs Class 2
(High Exposure/Poly Use vs High Exposure/Low or No Use)		 Panel C
Class 3 vs Class 1
(High Exposure/Poly Use vs No Exposure/Low or No Use)
Relative Risk Ratio [RRR] (95% Confidence Interval [CI])
Community factors
Neighborhood support 1.32 (1.04–1.68) * 1.22 (0.95–1.56) 0.92 (0.70–1.22)
Perceived availability of community youth activities 0.94 (0.85–1.04) 0.94 (0.85–1.05) 1.00 (0.89–1.12)
Racial discrimination (≥1 vs none) 0.32 (0.20–0.51) *** 0.82 (0.50–1.32) 2.53 (1.49–4.32) **
Behavioral factor
Missing school (≥1 day) 0.52 (0.28–0.98) * 1.55 (0.91–2.64) 2.95 (1.51–5.76) **
Emotional factors
Depressive symptoms (PHQ-9) 1.02 (0.96–1.08) 1.09 (1.03–1.15) ** 1.06 (1.01–1.14) *
Anxiety symptoms (GAD-7) 0.96 (0.91–1.02) 0.99 (0.92–1.05) 1.02 (0.95–1.10)
Suicidal thoughts (yes vs no) 0.42 (0.20–0.88) * 0.56 (0.30–1.05) 1.33 (0.58–3.07)
Demographic factors
Age 0.88 (0.77–1.01) 1.08 (0.93–1.24) 1.22 (1.04–1.43) *
Female (vs male) 1.20 (0.77–1.87) 1.05 (0.64–1.70) 0.88 (0.52–1.49)
Latinx (vs Black) 0.89 (0.55–1.46) 0.61 (0.37–1.01) 0.69 (0.40–1.19)
Other race (vs Black) 1.84 (0.80–4.23) 0.08 (0.01–0.64) * 0.04 (0.01–0.35) **
Multiracial (vs Black) 1.18 (0.47–2.96) 1.43 (0.63–3.28) 1.22 (0.47–3.14)
Open in a new tab
*

p < .05,

**

p < .01,

***

p < .001

Notes: Two separate models were run with Class 1 as the referent class (Panel C) and Class 2 as the referent class (Panels A and B), respectively.

Discussion

This study sought to identify distinct socioecological classes of adolescent substance use and neighborhood exposure and to examine how community, behavioral, emotional, and demographic factors differed among these groups. Using latent class analysis, three unique classes were identified: a no neighborhood violence and drug activity exposure and low or no substance use group, a high exposure/low or no use group, and a high exposure/polysubstance using group. Adolescents in the high-exposure classes shared higher likelihood of racial discrimination and missing school relative to the low-exposure group, while suicidal ideation was uniquely associated with the high exposure/low or no use class. Depressive symptoms were higher among adolescents in the high exposure/polysubstance use class, whereas anxiety symptoms did not differ across classes. Alternatively, higher perceived neighborhood support was observed among adolescents in the no exposure/low or no use class compared to the high exposure/low or no use class. The results indicate that differences across latent classes may be primarily structured by exposure to neighborhood risk rather than by substance use behavior. Together, these findings suggest that substance use may not represent a distinct adaptation pathway but rather reflects variation in severity within a broader context of shared environmental risk.

Neighborhood support emerged as an important strength when comparing youth with no reported exposure to neighborhood violence or visible drug activity and low or no substance use to those with high reported exposure and low or no substance use. This suggests that perceived neighborhood support may be more closely related to differences in perceived neighborhood exposure than to differences in substance use behavior. This aligns with theories of community-level resilience, which conceptualize social support as a buffer against visible disorder and substance availability (Prince et al., 2019). However, neighborhood support did not differentiate the two high exposure classes, indicating that it did not explain heterogeneity in polysubstance use among youth reporting similar exposure to neighborhood violence and visible drug activity. he mental health findings further underscore that risk among youth in high-exposure contexts is not limited to substance use. Suicidal ideation was uniquely associated with the high-exposure/low or no use class, while depressive symptoms were higher among youth in the high-exposure/polysubstance use class. Recent national evidence among adolescents exposed to neighborhood violence similarly found that suicidal ideation was highly prevalent, with depressive symptoms emerging as the strongest correlate and perceived lack of school safety also associated with suicidal ideation (Nikolova et al., 2025). Together, these findings suggest that youth exposed to neighborhood violence and visible drug activity may need mental health and school-based support even when they do not report substance use.

Interestingly, missing school (school absences) and experiencing racial discrimination were associated with membership in the high-exposure classes but did not differentiate polysubstance-using and non-using youth within those contexts. This pattern suggests that missing school and racial discrimination reflect shared structural and environmental burden rather than factors that distinguish substance use behavior within high-exposure contexts. Similarly, missing school was associated with membership in both high-exposure classes but did not differentiate between polysubstance-using and non-using youth within those contexts. Experiences of racial discrimination also emerged as a consistent correlate of the high exposure classes but did not differentiate substance use behaviors within those contexts. This pattern suggests that racial discrimination may operate as a shared structural stressor embedded within high-risk environments, rather than a factor that independently drives substance use behavior. Prior work using nationally representative data found that perceived racial discrimination was independently associated with higher odds of adolescent polysubstance use even after adjusting for psychosocial stressors (Baiden et al., 2023). In contrast, the present study findings suggest that racial discrimination may be more indicative of shared structural exposure than of which adolescents engage in substance use.

These findings collectively suggest that substance use and polysubstance use specifically may not be a consistent signal of risk among adolescents exposed to neighborhood risk. Thus, “low substance use” should not be interpreted as low need; prevention efforts in high-exposure communities should screen for mental health distress, suicidal ideation, discrimination-related stress, and school disengagement even among youth who are not reporting substance use.

4.1. Implications

These distinctions have significant implications for research, practice and policy. The results of the present study support multi-tiered approaches rather than substance-use-only programming. In this study, the high-exposure/low-use group may be especially important for early intervention because they would be easy to overlook if screening is triggered only by substance use behavior. Addressing this requires interventions that reduce exposure to neighborhood risk while also identifying youth experiencing distress regardless of whether they report substance use. For example, community-based interventions that strengthen neighborhood cohesion, invest in safe spaces, and promote youth civic engagement have demonstrated reductions in violence and related behavioral outcomes (Butts, Roman, et al., 2015; Butts, Wolff, et al., 2015). Place-based investments focused on improving neighborhood aesthetics have also been shown to reduce violence and improve mental health in urban communities (Branas et al., 2018). School-based strategies that target absenteeism have been associated with improvements in student well-being, particularly in high-risk settings (Durlak et al., 2011; Overstreet & Chafouleas, 2016). These approaches may serve as entry points for identifying youth in high-risk environments regardless of substance use status. Given that racial discrimination and violence co-occur as structural stressors in the present study, interventions should not be limited to individual-level behavior change but should also address the systems and community conditions that shape youth exposure to risk. This is particularly important because prior evidence has demonstrated mistrust of health services among youth in urban environments (Webb Hooper et al., 2019). Together, these findings support the use of community-based participatory research (CBPR) approaches in which community members help define priorities, design intervention strategies, and determine how services are delivered (Israel et al., 2001; Israel et al., 1998). In particular, youth participatory action research (YPAR) is an approach where young people are positioned as co-researchers who identify issues affecting their lives, collect and interpret data, and use findings to advocate for changes in their communities, and has been found to improve mental health and resilience among youth (Anyon et al., 2018; Cammarota & Fine, 2010; Raanaas et al., 2020; Smith et al., 2024). Co-developed approaches such as these may be especially important in contexts where racism, violence exposure, and institutional mistrust intersect (Woods-Jaeger et al., 2023). Additionally, these findings point to the importance of broader community-level policies that address systemic inequities, rather than intervention programs alone, as policies such as enforcement of anti-discrimination protections and investments into historically disinvested neighborhoods can improve mental health and reduce substance use disparities (Alegría et al., 2018; Williams et al., 2019). Recognizing that adaptation to environmental stressors is multifaceted and shaped by structural conditions can inform more equitable and comprehensive strategies for promoting adolescent well-being in high-risk environments.

4.2. Limitations

Several limitations should be considered when interpreting these findings. First, the cross-sectional design limits determination of causality or directionality among the measured concepts. Second, all measures were self-reported which is sensitive to recall and social desirability bias. Third, although this study employed community-based sampling methods to capture adolescents who may be underrepresented in school-based surveys, the non-random sampling design limits generalizability to broader youth populations. Fourth, while the latent class model identified meaningful psychosocial subgroups, unmeasured contextual factors (i.e., family structure, peer influences, or housing instability) were not included and could further shape class membership or explain differential membership characteristics. For example, other unmeasured factors may account for the low substance use group such as possible limited peer exposure to substance use or a lack of interest in engaging in such behaviors. Additionally, measures such as missing school do not capture underlying reasons, such as structural barriers or family context; therefore, conclusions should be interpreted with caution and may not be attributed to specific mechanisms. Similarly, neighborhood support was measured using a single item assessing whether neighbors are willing to help which does not capture the full multidimensional construct of collective efficacy, which may be more informative. Complete-case estimation reduced the multinomial regression sample size. Retention differed across the latent profiles and excluded participants were slightly older, indicating the potential for selection bias. Consequently, the multinomial regression estimates may not fully represent all participants included in the LCA and should be interpreted cautiously. Further, the study was conducted within a single urban community with a distinct racial and socioeconomic composition; future research should replicate these findings across different settings. Finally, future studies should incorporate longitudinal or mixed method designs to better capture developmental trajectories and contextual narratives.

4.3. Conclusions

This study identified distinct classes of neighborhood risk and substance use among urban adolescents. Findings suggest that high risk exposure is associated with racial discrimination, school absenteeism, and mental health burden, even among youth reporting low or no substance use. These results highlight that lower substance use does not necessarily indicate lower need and underscore the importance of screening and prevention approaches that address mental health, school disengagement, discrimination, and community conditions alongside substance use.

Highlights.

  • Identified three profiles of neighborhood exposure and substance use among urban youth.

  • High-exposure classes shared higher racial discrimination and school absenteeism.

  • Suicidal ideation was uniquely associated with the high-exposure/low-use class.

  • Depressive symptoms distinguished high-exposure/polysubstance use from high-exposure/low use.

  • Neighborhood support was higher among youth in the low-exposure/low-use class.

Footnotes

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.

Conflict of Interest

Authors have no conflicts of interest to declare.

Declaration of interests

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Ijeoma Opara reports financial support was provided by National Institute on Drug Abuse. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. (CDC), C. f. D. C. a. P. (2025). 1991-2023 High School Youth Risk Behavior Survey Data. http://yrbs-explorer.services.cdc.gov/
  2. Alaniz ML. (1998). Alcohol availability and targeted advertising in racial/ethnic minority communities. Alcohol Health and Research World, 22(4), 286. [PMC free article] [PubMed] [Google Scholar]
  3. Alegría M, NeMoyer A, Falgàs Bagué I, Wang Y, & Alvarez K. (2018). Social determinants of mental health: where we are and where we need to go. Current Psychiatry Reports, 20(11), 95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Anona K, Onigu-Otite E, Conner C, & Saxena K. (2024). Trauma, Resilience, and Substance Use in Adolescents: A Review Article. Adolescent Psychiatry. [Google Scholar]
  5. Anyon Y, Bender K, Kennedy H, & Dechants J. (2018). A systematic review of youth participatory action research (YPAR) in the United States: Methodologies, youth outcomes, and future directions. Health education & behavior, 45(6), 865–878. [DOI] [PubMed] [Google Scholar]
  6. Baiden P, Onyeaka HK, Aneni K, Wood B, LaBrenz CA, Muoghalu C, Peoples JE, Szlyk HS, Gobodzo EC, & Baiden JF. (2023). Perceived racial discrimination and polysubstance use among racial/ethnic minority adolescents in the United States. Drug and alcohol dependence, 248, 109894. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Biernacki P, & Waldorf D. (1981). Snowball sampling: Problems and techniques of chain referral sampling. Sociological methods & research, 10(2), 141–163. [Google Scholar]
  8. Branas CC, South E, Kondo MC, Hohl BC, Bourgois P, Wiebe DJ, & MacDonald JM. (2018). Citywide cluster randomized trial to restore blighted vacant land and its effects on violence, crime, and fear. Proceedings of the National Academy of Sciences, 115(12), 2946–2951. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brener ND. (2024). Overview and methods for the Youth Risk Behavior Surveillance System—United States, 2023. MMWR supplements, 73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Burdick-Will J, Stein ML, & Grigg J. (2019). Danger on the way to school: Exposure to violent crime, public transportation, and absenteeism. Sociological Science, 6, 118–142. [Google Scholar]
  11. Butts JA, Roman CG, Bostwick L, & Porter JR. (2015). Cure violence: a public health model to reduce gun violence. Annual review of public health, 36(1), 39–53. [DOI] [PubMed] [Google Scholar]
  12. Butts JA, Wolff KT, Misshula E, & Delgado SA. (2015). Effectiveness of the cure violence model in New York City. [Google Scholar]
  13. Cammarota J, & Fine M. (2010). Youth participatory action research: A pedagogy for transformational resistance. In Revolutionizing education (pp. 9–20). Routledge. [Google Scholar]
  14. Danielsson P. (2021). Collective efficacy and violent crime in suburban housing estates. European Journal of Criminology, 18(3), 345–365. [Google Scholar]
  15. Durlak JA, Weissberg RP, Dymnicki AB, Taylor RD, & Schellinger KB. (2011). The impact of enhancing students’ social and emotional learning: A meta-analysis of school-based universal interventions. Child development, 82(1), 405–432. [DOI] [PubMed] [Google Scholar]
  16. Etikan I, Musa SA, & Alkassim RS. (2016). Comparison of convenience sampling and purposive sampling. American journal of theoretical and applied statistics, 5(1), 1–4. [Google Scholar]
  17. Finkelhor D, Turner HA, Shattuck A, & Hamby SL. (2015). Prevalence of childhood exposure to violence, crime, and abuse: Results from the national survey of children’s exposure to violence. JAMA pediatrics, 169(8), 746–754. [DOI] [PubMed] [Google Scholar]
  18. Fonseca-Pedrero E, Díez-Gómez A, Pérez-Albéniz A, Al-Halabí S, Lucas-Molina B, & Debbané M. (2023). Youth screening depression: Validation of the Patient Health Questionnaire-9 (PHQ-9) in a representative sample of adolescents. Psychiatry Research, 328, 115486. [DOI] [PubMed] [Google Scholar]
  19. Fornander MJ, & Kearney CA. (2019). Internalizing Symptoms as Predictors of School Absenteeism Severity at Multiple Levels: Ensemble and Classification and Regression Tree Analysis. Front Psychol, 10, 3079. 10.3389/fpsyg.2019.03079 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Graif C, Gladfelter AS, & Matthews SA. (2014). Urban poverty and neighborhood effects on crime: Incorporating spatial and network perspectives. Sociology compass, 8(9), 1140–1155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Guy SM, Smith GM, & Bentler PM. (1994). The influence of adolescent substance use and socialization on deviant behavior in young adulthood. Criminal Justice and Behavior, 21(2), 236–255. [Google Scholar]
  22. Hart KR, Swaby MK, Oniya A, Okoye E, Onumah N, Bowens D, & Jones E. (2025). The Relationship Between Experiencing Neighborhood Violence and Mental Health Outcomes Among High School Students in the United States, YRBS 2023. Psychology International, 7(4), 93. [Google Scholar]
  23. Heads AM, Glover AM, Castillo LG, Blozis S, Kim SY, & Ali S. (2021). Perceived discrimination and risk behaviors in African American students: the potential moderating roles of emotion regulation and ethnic socialization. Journal of racial and ethnic health disparities, 8(2), 494–506. [DOI] [PubMed] [Google Scholar]
  24. Hendron M, & Kearney CA. (2016). School climate and student absenteeism and internalizing and behavioral behavioral problems. Children & Schools, 38(2), 109–116. [Google Scholar]
  25. Israel B, Schulz A, Parker E, & Becker A. (2001). Community-based participatory research: policy recommendations for promoting a partnership approach in health research. Education for health, 14(2), 182–197. [DOI] [PubMed] [Google Scholar]
  26. Israel BA, Schulz AJ, Parker EA, & Becker AB. (1998). Review of community-based research: assessing partnership approaches to improve public health. Annual review of public health, 19(1), 173–202. [DOI] [PubMed] [Google Scholar]
  27. Kingsbury M, Clayborne Z, Colman I, & Kirkbride JB. (2020). The protective effect of neighbourhood social cohesion on adolescent mental health following stressful life events. Psychological Medicine, 50(8), 1292–1299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lambert SF, Copeland-Linder N, & Ialongo NS. (2008). Longitudinal associations between community violence exposure and suicidality. Journal of Adolescent Health, 43(4), 380–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lardier DT, Opara I, Lin Y, Roach E, Herrera A, Garcia-Reid P, & Reid RJ. (2021). A spatial analysis of alcohol outlet density type, abandoned properties, and police calls on aggravated assault rates in a northeastern US city. Substance Use & Misuse, 56(10), 1527–1535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lim E, Davis J, Choi SY, & Chen JJ. (2019). Effect of sociodemographics, health-related problems, and family structure on chronic absenteeism among children. Journal of School Health, 89(4), 308–318. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Löwe B, Decker O, Müller S, Brähler E, Schellberg D, Herzog W, & Herzberg PY. (2008). Validation and standardization of the Generalized Anxiety Disorder Screener (GAD-7) in the general population. Medical care, 46(3), 266–274. [DOI] [PubMed] [Google Scholar]
  32. Motyka MA, & Al-Imam A. (2022). Causes of drug initiation among adolescents. Canadian Journal of Family and Youth/Le Journal Canadien de Famille et de la Jeunesse, 14(1), 63–81. [Google Scholar]
  33. Nikolova S, Small E, & Sesay B. (2025). Suicidal ideation in US Adolescents exposed to neighborhood violence. Adolescents, 5(3), 31. [Google Scholar]
  34. Olivier E, Morin AJ, Langlois J, Tardif-Grenier K, & Archambault I. (2020). Internalizing and behavioral behavior problems and student engagement in elementary and secondary school students. Journal of Youth and Adolescence, 49(11), 2327–2346. [DOI] [PubMed] [Google Scholar]
  35. Opara I, Weerakoon SM, Brooks Stephens JR, Choe T, Gunn JF 3rd, & Thrasher SS. (2025, Feb). Relationship between suicide ideation and attempts, bully victimization, dating violence, and depressive symptoms among Black and Hispanic youth. Suicide Life Threat Behav, 55(1), e13015. 10.1111/sltb.13015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Opara I, Lardier DT, Metzger I, Herrera A, Franklin L, Garcia-Reid P, & Reid RJ. (2020). “Bullets have no names”: A qualitative exploration of community trauma among Black and Latinx youth. Journal of child and family studies, 29(8), 2117–2129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Opara I, Leonard NR, Thorpe D, & Kershaw T. (2021). Understanding Neighborhoods’ Impact on Youth Substance Use and Mental Health Outcomes in Paterson, New Jersey: Protocol for a Community-Based Participatory Research Study. JMIR Research Protocols, 10(5), e29427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Opara I, Weerakoon S, Saeed G, Duran-Becerra B, & Pagan C. (2026). Qualitative Exploration of Neighborhoods, Youth Substance Misuse, and Mental Health in an Urban Community. Journal of Clinical Child & Adolescent Psychology, 1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Overstreet S, & Chafouleas SM. (2016). Trauma-informed schools: Introduction to the special issue. School mental health, 8(1), 1–6. [Google Scholar]
  40. Perreira KM, Marchante AN, Schwartz SJ, Isasi CR, Carnethon MR, Corliss HL, Kaplan RC, Santisteban DA, Vidot DC, & Van Horn L. (2019). Stress and resilience: Key correlates of mental health and substance use in the Hispanic community health study of Latino youth. Journal of immigrant and minority health, 21(1), 4–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Pinderhughes H, Davis R, & Williams M. (2015). Adverse community experiences and resilience: A framework for addressing and preventing community trauma. [Google Scholar]
  42. Prince DM, Hohl B, Hunter BA, Thompson AB, Matlin SL, Hausman AJ, & Tebes JK. (2019). Collective efficacy as a key context in neighborhood support for urban youth. American journal of community psychology, 63(1-2), 179–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Qeadan F, Azagba S, Barbeau WA, Gu LY, Mensah NA, Komaromy M, English K, & Madden EF. (2022). Associations between discrimination and substance use among college students in the United States from 2015 to 2019. Addictive behaviors, 125, 107164. [DOI] [PubMed] [Google Scholar]
  44. Raanaas RK, Bjøntegaard H, & Shaw L. (2020). A scoping review of participatory action research to promote mental health and resilience in youth and adolescents. Adolescent Research Review, 5(2), 137–152. [Google Scholar]
  45. Smith KE, Acevedo-Duran R, Lovell JL, Castillo AV, & Cardenas Pacheco V. (2024). Youth are the experts! Youth participatory action research to address the adolescent mental health crisis. Healthcare, [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. US Census Bureau - QuickFacts Paterson, NJ. (2021). https://www.census.gov/quickfacts/fact/table/NJ,patersoncitynewjersey/PST045221
  47. Verdery AM, Weir S, Reynolds Z, Mulholland G, & Edwards JK. (2019). Estimating hidden population sizes with venue-based sampling: extensions of the generalized network scale-up estimator. Epidemiology, 30(6), 901–910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Vermeiren R, Ruchkin V, Leckman PE, Deboutte D, & Schwab-Stone M. (2002). Exposure to violence and suicide risk in adolescents: A community study. Journal of Abnormal Child Psychology, 30(5), 529–537. [DOI] [PubMed] [Google Scholar]
  49. Wang D, Choi J-K, & Shin J. (2020). Long-term neighborhood effects on adolescent outcomes: mediated through adverse childhood experiences and parenting stress. Journal of Youth and Adolescence, 49(10), 2160–2173. [DOI] [PubMed] [Google Scholar]
  50. Webb Hooper M, Mitchell C, Marshall VJ, Cheatham C, Austin K, Sanders K, Krishnamurthi S, & Grafton LL. (2019). Understanding multilevel factors related to urban community trust in healthcare and research. International journal of environmental research and public health, 16(18), 3280. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Weerakoon SM, Henson-Garcia M, Abraham A, Vidot DC, Messiah SE, & Opara I. (2023). Adolescent polysubstance use and co-occurring weapon carrying, bullying victimization, and depressive symptomology: patterns and differences in the United States. Child Psychiatry & Human Development, 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Williams DR, Lawrence JA, Davis BA, & Vu C. (2019). Understanding how discrimination can affect health. Health services research, 54, 1374–1388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Woods-Jaeger B, Knutzen KE, Lucas D, Cave N, Latimer SK, Adams H, Abreon Bates, H., Aceyeah, & Renfro TL. (2023). Anti-racist violence prevention: Partnering with Black youth to identify intervention priorities. Health Promotion Practice, 24(2), 223–231. [DOI] [PubMed] [Google Scholar]
  54. Zimmerman GM, & Messner SF. (2013). Individual, family background, and contextual explanations of racial and ethnic disparities in youths’ exposure to violence. American Journal of Public Health, 103(3), 435–442. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES