Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Feb 1.
Published in final edited form as: Clin Psychol Rev. 2022 Nov 30;99:102233. doi: 10.1016/j.cpr.2022.102233

Culturally Sensitive Prevention Programs for Substance Use Among Adolescents of Color: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Ai Bo 1,*, Trenette Clark Goings 2, Caroline B R Evans 2, Anjalee Sharma 2, Zoe Jennings 2, Brenna Durand 3, Angela Bardeen 4, Andrea Murray-Lichtman 2
PMCID: PMC9847495  NIHMSID: NIHMS1858800  PMID: 36495737

Abstract

This systematic review and meta-analysis aims to synthesize the efficacy of culturally sensitive prevention programs for substance use outcomes among U.S. adolescents of color (aged 11 to 18 years old) and explore whether the intervention effects vary by participant and intervention characteristics. Eight electronic databases and grey literature were searched for eligible randomized controlled trials through September 2022. Robust variance estimation in meta-regression was used to synthesize treatment effect size estimates and to conduct moderator analysis. After screening, 30 unique studies were included. The average treatment effect size across all substance use outcomes (including 221 effect sizes) was Hedges’s g = −0.20, 95% CI = [−0.24, −0.16]. The synthesized effect sizes were statistically significant across types of substances (alcohol, cigarette, marijuana, illicit and other drugs, and unspecified substance use), racial/ethnic groups (Hispanic, Black, and Native American), and at different follow-ups (0–12 months, more than 12 months). Very few studies reported substance use consequences as outcomes and the synthesized effect size was non-significant. Meta-regression findings suggest that the intervention effect may vary based on the type of substance. This meta-analysis found supportive evidence of culturally sensitive prevention programs’ efficacy in preventing or reducing substance use among Black, Hispanic, and Native American adolescents. More substance use prevention efforts and evidence is needed for Asian American, Pacific Islander, and multiracial adolescents.

Keywords: substance use prevention, youth, cultural adaptation, alcohol, cigarette smoking, marijuana

Introduction

Adolescent Substance Use Prevalence, Consequences, and Racial/Ethnic Disparities

Adolescent substance use is a pressing yet preventable public health problem associated with various negative developmental, behavioral, health, mental health, and social consequences (Carbia et al., 2018; Moore et al., 2007; Scheier & Griffin, 2021). According to the Monitoring the Future study (MTF; Miech et al., 2021), lifetime prevalence of substance use increases substantially from 8th to 12th grade for alcohol (26% and 62%), cigarettes (11.5% and 24%), marijuana (15% and 44%), nicotine vaping (23% and 44%), and any illicit drug use including marijuana (21% and 47%). Racial/ethnic disparities exist in adolescent substance use prevalence, trajectories, and consequences. For instance, according to the 2019 High School Youth Risk Behavior Survey, Native American, multiracial, Black, and Hispanic adolescents reported higher rates of initiating cigarette smoking, alcohol use, and marijuana use before age 13 than White and Asian American adolescents (Centers for Disease Control and Prevention, 2019). Data based on the 2015 to 2019 National Surveys on Drug Use and Health indicated that Native American and multiracial adolescents had higher rates of past-year marijuana use and illicit drug use than Hispanic, Black, White, and Asian American adolescents (Center for Behavioral Health Statistics and Quality, 2021). MTF study found that Hispanic 8th and 10th graders had higher rates of past 30-day binge drinking and past 12-month illicit drug and marijuana use than Black and White adolescents (Miech et al., 2021). These data highlight that Native American, multiracial, Hispanic, and Black adolescents are at increased risk for early substance use initiation and certain types of substance use.

Adolescents of color who use substances are at higher risk for negative consequences than White youth who use substances throughout the life course (Chartier & Caetano, 2010; Wallace, 1999; Zapolski et al., 2014). Black youth are disproportionately arrested for substance use and receive more severe dispositions at each juncture of court processing than White youth (Belenko et al., 2004). Related, Hispanic adolescent drug offenders received longer custody in state facilities than Black and White youth (306 days, 235 days, and 144 days, respectively; Poe-Yamagata & Jones, 2000). Black and Hispanic youth with substance use disorders are less likely to receive adequate treatment compared to White adolescents (Alegria et al., 2011). These disparities and inequities continue into adulthood. Black and Hispanic youth are less likely to stop heavy drinking as they transition to adulthood and suffer from high rates of problem drinking, negative social consequences, and alcohol dependence (Mulia et al., 2008; 2017). Among inpatient substance use treatment seekers, Black individuals reported significant delays in treatment entry (about 4 to 5 years) relative to White individuals, regardless of socioeconomic status (Lewis et al., 2018). Native American individuals aged 12 years or older had the highest rates of past-year alcohol and substance use disorders among monoracial groups (Center for Behavioral Health Statistics and Quality, 2021) and almost three times higher rate of alcohol-attributable death than White individuals (Landen et al., 2014). Asian American individuals had the lowest rates of substance use treatment utilization among people who needed treatment across all racial/ethnic groups (Center for Behavioral Health Statistics and Quality, 2021). This research highlights the substantial racial/ethnic disparities related to substance use consequences that must be addressed by early, culturally sensitive prevention efforts.

Factors Influencing Substance Use Among Adolescents of Color

Researchers have identified many factors that impact the likelihood of adolescent substance use across racial/ethnic groups, including individual factors (e.g., substance use expectancies, sensation seeking, mental health problems; Bardo et al., 2007; Hussong et al., 2017; Montes et al., 2019), peer influences (e.g., peer substance use behaviors and norms; Leung et al., 2014), family and parental factors (e.g., substance use among parents, parent-child relationships, parental monitoring and support, parental attitudes, rules and monitoring of teen substance use, family conflict; Sharmin et al., 2017; Tael‐Öeren et al., 2019; Yap et al., 2017), and various community factors (e.g., community-level availability of and exposure to alcohol and other substances, social/communal norms about substance use; Hawkins et al., 2004). In addition, research suggests that certain risk and protective factors unique to adolescents of color influence their substance use. For example, experiences of racial/ethnic discrimination (Benner et al., 2018; Cave et al., 2020), higher levels of acculturation (i.e., greater English speaking proficiency, longer residence in the U.S., and second versus first generation of immigrants; Sirin et al., 2022), and higher parent–youth acculturation discrepancy (Nair et al., 2018; Unger et al., 2009) are associated with increased risk of substance use. Cultural socialization (i.e., the process by which parents share messages to their children regarding cultural heritage, pride, values, and practices; Ayón et al., 2020; Grindal & Nieri, 2016; Neblett et al., 2010; Nieri et al., 2020; Zapolski & Clifton, 2019) and a positive sense of one’s racial/ethnic identity may protect against adolescent substance use (Fisher et al., 2017; Sanchez et al., 2018; Unger et al., 2020; Zapolski et al., 2017). Culturally sensitive substance use prevention programs tend to address many of the aforementioned factors that influence substance use among adolescents of color.

Overview of Culturally Sensitive Substance Use Prevention Interventions

Culturally sensitive substance use prevention programs usually address a combination of positive youth development factors (e.g., future orientation, problem solving, communication, decision making, emotional regulation), substance use-specific factors (e.g., substance use knowledge, expectancies, peer norms, resistant skills), and racially/ethnically-specific factors (e.g., racial/ethnic socialization, racial/ethnic identity, acculturation) that are related to adolescent substance use. Culturally sensitive prevention programs addressing these factors were embedded within family, school, or community contexts to deliver the intervention. The intervention components, participants, delivery approach, interventionist, and setting necessarily varied depending on the targeted contexts. For example, programs targeting the family context focus on parents as the agents of change, address various protective parenting behaviors and parental racial/ethnic socialization practices, can involve parents or adult primary caregivers with or without adolescent participation (Bo et al., 2018). Parent-based programs can be self-guided online or at home, or delivered by professionals in various settings (e.g., at home, in a group format in a school or community setting, or online). Culturally sensitive programs implemented in school contexts generally address positive youth development, peer influences on substance use, and racial/ethnic identity. School-based programs can be delivered by teachers in class, or delivered by community members or peer leaders in a group format, and may or may not involve the participation of parents or adult primary caregivers (Onrust et al., 2016). Programs embedded in community contexts can be led by community members or professionals and conducted in various formats (e.g., group discussion, workshop, community gathering, community awareness campaign).

Culturally sensitive prevention programs may target adolescents in one particular racial/ethnic group or adolescents across racial/ethnic groups. Cultural tailoring of interventions occurs along a continuum from culturally adapted programs (i.e., modifying an existing empirically supported program) to culturally grounded programs (i.e., originally designed with a specific culture in mind so that the program reflects the values, behaviors, and norms of the target population; Okamoto et al., 2014). Cultural adaptation processes range from surface-structure adaptations (e.g., minor modifications to images and terms used in the curriculum) to deep-structure adaptations (e.g., intertwining the prevention program components with cultural phenomena). This systematic review focuses on deep structure-adapted or culturally grounded programs, which we termed culturally sensitive programs, consistent with prior relevant literature.

A few systematic reviews and meta-analyses have examined the effects of culturally sensitive programs on adolescent substance use. Notably, findings of previous meta-analyses generally indicate small or non-significant effect sizes. One meta-analysis of 10 studies (half of which used a quasi-experimental design) found small effects of culturally sensitive interventions on recent alcohol use (g = 0.225, p < .05), while effects on recent marijuana use were non-significant (Hodge et al., 2012). Another meta-analysis of 10 quasi-experimental and experimental studies of culturally adapted substance use prevention and interventions programs for Hispanic adolescents found trivial effects (g = 0.06, p < .05) at posttest and small effects (g = 0.26, p < .05) at follow-up on substance use outcomes (Hernandez Robles et al., 2018).

Importance of Current Review

The racial/ethnic disparities in substance use rates and consequences warrant rigorous research to synthesize the most up-to-date evidence of the effects of culturally sensitive prevention programs on substance use for various groups of adolescents of color. Most of the previous systematic reviews on culturally sensitive substance use prevention included a mix of experimental and quasi-experimental designs, focused on only one racial/ethnic group, and lacked (or had limited) moderator analysis. The current review fills these research gaps in two respects. First, we conducted a comprehensive review of culturally sensitive prevention programs evaluated in randomized control trials that examined various substance use outcomes among U.S. adolescents of color between the ages of 11 and 18 years. Focusing only on RCTs provides more robust evidence of the efficacy of the programs in question. Including all racial/ethnic groups in one review enabled us to identify subgroup differences in intervention effects and identify racial/ethnic groups for which more culturally sensitive prevention efforts are needed. Second, we conducted moderator analyses to explore whether the treatment effects vary by types of substance use outcomes, participant characteristics (e.g., age, sex, race/ethnicity), intervention characteristics (e.g., main intervention content, delivery methods), control conditions, assessment timings, and risk of bias ratings. Understanding the moderators is critical for refining existing prevention programs to maximize the treatment effects for subgroups of adolescents. The study protocol was registered on PROSPERO (CRD42020215500; Bo et al., 2020).

Methods

Inclusion and Exclusion Criteria

Included types of studies: Randomized controlled trials (individual or cluster design).

Included types of participants: U.S. Black, Hispanic, Native American, Asian American, Native Hawaiian or Pacific Islander, and multiracial adolescents between the ages of 11 and 18 years.

Included types of interventions: Any culturally sensitive universal, selective, or indicated individual-, family-, school-, or community-based substance use prevention programs or positive youth development programs conducted in the United States. Culturally sensitive program components include culturally appropriate language or contents that address the targeted racial and ethnic groups’ cultural values, racial and ethnic identity, pride, acculturation, discrimination, or racial/ethnic socialization. Eligible prevention programs could be deep-structure adapted from an effective substance use prevention program or culturally grounded (i.e., designed explicitly for adolescents of color).

Included types of outcomes: Any behavioral substance use outcomes and substance use related consequences or problems. Behavioral substance use outcomes include ever used substance during one’s lifetime, future substance use intentions, and substance use frequency and amount within any timeframe (e.g., past month, past three months). All kinds of substances were eligible (e.g., alcohol, tobacco, marijuana, cocaine, inhalants, prescription medication, over-the-counter medication, other illicit drugs, and unspecified substance use).

Included types of comparisons: No treatment, waitlist control, treatment as usual, attention control, or alternative intervention with a similar format to the intervention group (e.g., non-culturally sensitive prevention).

Our systematic review excluded studies that had adolescent participants with a substance or alcohol abuse and dependence diagnosis or graduated from high school; studies that mainly had White participants; studies that focused on prevention programs that were not culturally sensitive or that were surface-structure adapted (e.g., only used culturally relevant language); studies that only used passive intervention messages (e.g., showing posters that have anti-substance use messages) without participant interaction or actively learning new skills; studies that did not report substance use behavioral outcomes (i.e., only reported substance use attitudes, expectancies, resistant skills); or studies that did not report adequate information for effect size calculation.

Search Methods for Identifying Studies

We followed the Cochrane recommendations for searching bibliographic databases, trial registries, and grey literature (Higgins et al., 2020). A university librarian with experience conducting systematic reviews searched the following electronic bibliographic databases in September 2020, and the first author updated the search in August 2021 and September 2022. The databases are the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE (via PubMed), PsycINFO (via EBSCO), Embase, CINAHL, Social Services Abstracts, Web of Science, and ProQuest Dissertations and Theses. In addition, the first author searched ClinicalTrials.gov and several evidence-based practice registries to identify relevant studies (i.e., What Works Clearinghouse, Blueprints, and Office of Juvenile Justice and Delinquency Prevention Model Program Guide). Studies that cited candidate articles and reference lists of candidate articles and relevant systematic reviews were also searched. Searching strategies (e.g., the use of controlled vocabulary) varied slightly for different databases based on the following key search terms: (“African American” OR Asian* OR black OR ethnic OR ethnicity OR Hispanic* OR Latina OR Latino OR Latinx OR indigenous OR “Native American” OR “American Indian” OR “Pacific Islander” OR race OR racial OR minority OR minorities OR immigrant*) AND (youth OR “young people” OR adolescen* OR teen* OR student OR child* OR girls OR boys OR “school age”) AND (substance OR alcohol OR drug OR drinking OR tobacco OR smoking OR cigarette OR vaping OR vape OR marijuana OR opioid OR inhalant OR stimulants) AND (intervention OR outcome OR trial OR experiment* OR evaluation OR program OR prevention OR preventive OR services OR “positive youth development” OR education OR training OR management OR curriculum) AND (cultur* OR tailor* OR adapt* OR sensitive OR adjust OR multicultural OR “cross-cultural” OR bicultural OR intercultural) AND (“randomized controlled trial” OR “controlled clinical trial” OR random* OR trial* OR “control group” OR “comparison group”). Complete search strategies are available upon request.

Data Collection and Analysis

Selection of Studies

All reviewers were trained in the current study’s systematic review protocol. Two reviewers independently screened the titles and abstracts for all the records based on the inclusion and exclusion criteria. After excluding irrelevant records, the two reviewers independently reviewed the full text of the remaining records. The screening was conducted on Covidence.org (Covidence, n.d.). Disagreements between the two reviewers were resolved through discussion with a third reviewer. The average agreement (i.e., number of citations with a consensus in a reviewer pair divided by the total number of citations screened) was 92% for the title and abstract screening and 80% for the full-text review, which is satisfactory.

Data Extraction and Management

A data extraction sheet based on the Cochrane data collection form for intervention reviews (RCTs only) was created and added to Covidence.org (Higgins & Green, 2011). Two reviewers independently extracted data on study characteristics and effect size. A third reviewer compared the extracted data by the two reviewers and resolved any disagreements. Information extracted included study source (i.e., title, author, document type [e.g., dissertation, report, publication], year of publication), study design (i.e., RCT, cluster RCT), inclusion and exclusion criteria for participation in each study, sample characteristics (i.e., age, sex, race and ethnicity), intervention characteristics (i.e., components, duration, frequency, setting, intervention provider, fidelity), control condition, outcome measures, and assessment or follow-up timing. Extracted data related to effect size included sample size, means and standard deviations for continuous outcomes, a two-by-two table for dichotomous outcomes, or effect estimates (e.g., odds ratios, regression coefficients) with confidence intervals or standard errors. We extracted effect size data at all available time points (e.g., post-tests, follow-ups). The data collection sheet was pilot tested with a few targeted articles and was modified before formal data extraction.

Assessment of Risk of Bias in the Included Studies and Publication Bias

Two reviewers independently assessed the risk of bias for the included studies based on the Cochrane Risk of Bias tool 2 (RoB 2; Sterne et al., 2019). RoB 2 consists of five domains: (1) risk of bias arising from the randomization process, (2) risk of bias due to deviations from the intended interventions, (3) risk of bias due to missing outcome data, (4) risk of bias in measurement of the outcome, and (5) risk of bias in the selection of the reported result. Each domain was rated as having a low risk of bias, some concerns, or a high risk of bias with an explanation based on several signaling questions. Regarding the overall risk of bias, a study was judged to have a high overall risk of bias if at least one domain was judged as having a high risk of bias. A study was judged to raise some concerns if at least one domain had some concerns and no domain was judged as having a high risk of bias. Risk of bias assessment findings were incorporated into the meta-analyses using the five individual domains (as well as the overall risk of bias) as independent variables in meta-regression models. This approach provides specific information regarding which risk of bias domain influences the effect size estimates (Bo et al., 2021). The funnel plot trim-and-fill method (Duval & Tweedie, 2000) was used to explore possible bias due to missing results in the overall effect size by accounting for effect sizes from the estimated number of missing studies.

Data Synthesis

All effect sizes were calculated and converted to Hedges’s g with standard errors (Hedges & Olkin, 1985). Negative effect sizes indicate a reduction in substance use outcomes. Because multiple effect sizes of substance use outcomes were extracted from the same study and these effect sizes are not statistically independent, robust variance estimation in meta-regression was used to estimate overall treatment effects and to conduct moderator analysis while adjusting for the dependencies among the effect sizes (Hedges et al., 2010; Tanner-Smith & Tipton, 2014). The robust meta-regression method assumes a pre-specified level of dependency as expressed in the form of a correlation (Hedges et al., 2010). Sensitivity analyses were conducted for average dependency correlations ranging from 0.10 to 0.90 and 0.8 was used in the final models reported in this paper. The robumeta package in Stata 17 (StataCorp, 2019) was used to perform the meta-analysis.

Moderator analyses were conducted to explore whether the effects of culturally sensitive prevention programs for substance use outcomes varied across a priori defined moderators, including participants’ characteristics (i.e., mean age, sex or the percentage of female, and race/ethnicity), intervention characteristics (i.e., duration of the intervention, setting, and intervention type: parent-based versus non-parent-based, addressed substance use specific factors versus positive youth development only, culturally grounded versus culturally adapted), control conditions (i.e., non-active versus active controls), assessment timing (0–12 months versus more than 12 months), type of substance use outcomes, and ratings for individual risk of bias domains as well as the overall risk of bias. Moderator analyses are likely to be less powerful than tests for the average effect size (Hedges & Pigott, 2004). Therefore, power analysis (using the metapower package in R) was conducted to determine if the power is sufficiently high to detect meaningful differences in effects across subgroups.

Sensitivity analyses investigated whether the effects of culturally sensitive prevention programs on substance use outcomes remained the same after removing outliers. If the effects were comparable before and after the sensitivity analysis, the results were robust to outliers. We considered an effect size as an outlier if the effect size’s confidence interval (CI) did not overlap with the CI of the pooled effect (i.e., if the upper bound of the effect size’s 95% CI was lower than the lower bound of the pooled effect size’s CI or the lower bound of the effect size’s CI was higher than the upper bound of the pooled effect size CI; Viechtbauer & Cheung, 2010).

Results

Searches of the eight electronic bibliographic databases yielded a total of 2473 records in September 2020, and the updated searches in August 2021 and September 2022 yielded an additional 131 and 196 records, respectively. A grey literature search yielded a total of 184 records. After removing 1273 duplicates, 1711 records were included for the title and abstract screening and 152 of these records were further included for the full-text review. Finally, 36 reports representing 30 unique studies were included in the systematic review and meta-analysis. Figure 1 illustrates the search and screening process. Appendix A presents the intervention and participant characteristics of the 30 included studies. Appendix B lists the references of the included studies.

Figure 1.

Figure 1

Open in a new tab

Flow Diagram for Study Selection Process

Study Characteristics

Table 1 summarizes the key intervention and participant characteristics of the included studies. Among the 30 included RCTs, 20 studies randomized participants at the individual level and 10 randomized at the school or county level. All studies were conducted in the United States. Regarding the types of interventions, 16 studies involved substantial participation by parents or adult primary caregivers (i.e., 8 parent-based interventions, 5 combined parent and adolescent programs, 3 separate adolescent and parent interventions), 14 studies targeted adolescents (i.e., 4 in-class activities for students, 6 adolescent groups or workshops, 1 brief intervention, 2 online self-paced programs, and 1 school-based campaign). Among the 30 included studies, 23 studies addressed both substance use-specific factors (e.g., substance use knowledge, attitudes, norms, resistance skills, substance use specific parenting) and positive youth development (e.g., social and communication skills, assertiveness, decision making, problem solving, self-efficacy, emotional regulation, goal setting), 4 studies focused only on positive youth development, and 3 studies focused only on substance use-specific factors. Eleven studies used interventions that can be categorized as culturally adapted whereas nineteen studies used interventions that can be categorized as culturally grounded. Sixteen interventions were implemented in school settings, 11 interventions were implemented in community settings, and 3 interventions were implemented entirely online. The average duration of the interventions was about 20 hours (SD = 16). Regarding the control condition, four studies had active control groups only (i.e., two studies used substance use prevention programs without cultural adaptation, two studies used youth program only as control condition when compared with combined parent-youth programs as the intervention). Twenty-four studies had non-active control groups only (i.e., 17 had treatment as usual or attention controls that addressed no or minimal substance use-specific factors, and 7 had no treatment or waitlist control conditions). Two studies had both active and non-active controls.

Table 1.

Summary of Intervention and Participant Characteristics

Study Characteristics N
Design
Individual RCT 20
Cluster RCT 10
Intervention type
Targeting Hispanic adolescents 16
Parent groups
 - Familias Unidas: 6 studies
 - Parent Management Training: 2 studies		 8
Combined parent and adolescent groups
 - Bridges to High School: 1 study		 1
Separate adolescent and parent intervention
 - keepin’ it REAL + New Generation: 2 studies
 - Modified Project Towards No Tobacco Use + Raising Smoke-Free Kids: 1 study		 3
Adolescent groups or workshops
 - Sembrando Salud: 1 study		 1
In class activities for students
 - keepin’ it REAL (kiR): 1 study
 - Project Fun Learning About Vitality, Origins, and Respect (FLAVOR): 1 study		 2
Online self-paced program for adolescents
 - Vamos: 1 study		 1
Targeting Black adolescents 7
Combined parent and adolescent groups
 - Strong African American Families: 3 studies
 - Adults in the Making: 1 study		 4
Adolescent groups or workshops
 - Promoting Health Among Teens (PHAT): 1 study
 - Focus on Kids: 1 study		 2
Brief intervention for adolescents
 - Project Socialization and Activation of Feedback Exchange: 1 study		 1
Targeting Native American adolescents 7
In class activities for students
 - Living in 2 Worlds: 1 study
 - Circle of Life (CoL): 1 study		 2
Adolescent groups or workshops
 - Respecting the Circle of Life: 1 study
 - Motivational Interviewing and Cultural for Urban Native American Youth + Cultural Wellness Gathering: 1 study
 - Arrowhead Business Group intervention: 1 study		 3
School-based drug prevention campaign
 - Be Under Your Own Influence – American Indian (BUYOI‐AI): 1 study		 1
Online self-paced program for adolescents
 - SmokingZine: 1 study		 1
Main intervention component
Substance use prevention 3
Positive youth development 4
Both substance use prevention and positive youth development 23
Type of culturally sensitive programs
Culturally grounded 19
Culturally adapted 11
Setting
School 16
Community 11
Online 3
Duration (hour) Mean = 20, SD = 16
Control
Active: Alternative interventions
 - Substance use prevention without cultural adaptation: 2 studies
 - A control group with only youth participants when compared with a parent-youth treatment group: 2 studies		 4
Non-Active:
 - Treatment as usual or attention control: 17 studies
 - No treatment or waitlist control: 7 studies		 24
Both active and non-active 2
Adolescent participant sex (% female) Mean = 51%, SD = 7%
Adolescent participant age (year) Mean = 13.6, SD = 1.6
Open in a new tab

Participant Characteristics

The sample size of the included comparisons of each treatment group to a control group ranged from 31 to 2212. The average percentage of adolescent female participants was 51% (SD = 7%). The average age of adolescent participants was 13.6 years (SD = 1.6). Sixteen studies focused on or had a majority of Hispanic adolescents (mainly from urban settings), 7 on Black adolescents (4 rural, 3 urban), and 7 on Native American adolescents (5 reservations, 2 urban). Most studies (27 out of 30) targeted the general adolescent population while three studies targeted adolescents at higher risk for substance use (e.g., adolescents with behavior problems such as disruptive behaviors, destruction of property, harmful or offensive behaviors or having been arrested).

Summary of Culturally Sensitive Components

Among interventions that targeted Hispanic adolescents, most of them addressed acculturation issues (e.g., adolescents’ bicultural competency, parental knowledge of U.S. cultural norms), Hispanic cultural values (e.g., familismo/family orientation, respeto/respect, personalismo/personal treatment, simpatía/sympathy) and Hispanic ethnic identity (e.g., by using Telenovela) through cultural socialization. Only a few studies addressed racial and ethnic discrimination and immigration-related challenges. Among interventions that targeted Black adolescents, most of the culturally sensitive components focused on two types of ethnic and racial socialization: cultural socialization (e.g., building positive racial/ethnic and cultural pride) and preparation for racial bias (e.g., strategies of handling discrimination, culturally appropriate coping strategies). Among interventions that targeted Native American adolescents, culturally sensitive components mainly focused on cultural socialization, including using Native American cultural symbols (e.g., medicine wheel), cultural values, knowledge, stories, illustrations, traditional practices, encouraging adolescents to learn about tribal-specific histories, and promoting ethnic and cultural identity.

Risk of Bias Assessment

Figure 2 presents the risk of bias assessment results for the included studies based on RoB 2. Ratings for individual studies are presented in Appendix C. Most of the studies (90%) were rated as having some concerns and a few studies (10%) were rated as high risk in the randomization process because they did not clearly report information on random sequence generation, allocation concealment, or baseline balance. About half of the studies (52%) were rated as having a low risk of bias due to deviations from intended interventions. About a third of the included studies (32%) were determined to have some concerns about potential deviation and 16% of the studies were rated high risk in terms of deviations from the intended intervention, mainly because blinding or masking participants’ and providers’ knowledge about treatment assignments is often infeasible for behavioral interventions and intention-to-treat analyses were not used or not properly used. About a third of the studies (32%) were judged to have a low risk of bias due to missing outcome data because they either had low attrition (< 5%) or demonstrated evidence that the result was not biased by missing outcome data. In contrast, 29% of the included studies generated some concerns and 39% of the included studies had a high risk of bias due to missing outcome data (i.e., had more than 5% attrition and did not show any evidence that the result was not biased by missing outcome data). The majority of the included studies (77%) were rated as having some concerns related to measurement of outcomes because they used self-report outcomes on surveys and did not have comparable control groups, and therefore outcome assessors were not blinded or masked to the intervention assignments. Several studies (16%) were rated as having a low risk of bias in the measurement of the outcome mostly because they had comparable control groups, and therefore the outcome assessment was less likely to be influenced by knowledge of participants’ intervention assignments. Nearly half of the included studies were rated as either having a high risk of bias (39%) or some concerns (7%) in their selection of reported results because there was insufficient information to determine whether data were analyzed per a pre-specified analysis plan. Finally, when all domains were collectively considered, 39% of the studies were judged to have some concerns overall and 61% of the studies were judged to have a high risk of overall bias.

Figure 2.

Figure 2

Open in a new tab

Percent of Studies Exhibiting Degrees of Bias

Effects of Intervention

Overall Effect

Table 2 presents the effect size estimates overall and for each substance use outcome (i.e., alcohol use, cigarette smoking, marijuana use, illicit and other drug use, unspecified substance use, and substance use consequences). The overall mean treatment effect size for all reported substance use outcomes, with 221 effect sizes from 30 studies, was g = −0.20 (p < .05) with a 95% CI = [−0.24, −0.16]. The synthesized effect sizes for all substance use outcomes (except for substance use consequences) and both assessment or follow-up timings (0–12 months, more than 12 months) were statistically significant. Effect sizes were also synthesized for each racial and ethnic group. The mean treatment effect size was statistically significant among programs targeting Hispanic adolescents (g = −0.21, 95% CI = [−0.26, −0.16]), Black adolescents (g = −0.24, 95% CI = [−0.37, −0.11]), and Native American adolescents (g = −0.13, 95% CI = [−0.23, −0.02]).

Table 2.

Effect Size Estimate for Substance Use Outcomes

Original Sensitivity
K1 K2 Hedges’s g 95% CI K1 K2 Hedges’s g 95% CI
Overall 221 30 −0.20* −0.24, −0.16 203 30 −0.18* −0.21, −0.15
Outcome
 Alcohol use 61 21 −0.14* −0.20, −0.09 56 21 −0.13* −0.18, −0.08
 Cigarette smoking 51 17 −0.21* −0.32, −0.10 45 17 −0.20* −0.30, −0.11
 Marijuana use 36 13 −0.14* −0.22, −0.06 31 13 −0.15* −0.22, −0.09
 Illicit and other drug use a 28 8 −0.32* −0.50, −0.14 27 7 −0.27* −0.41, −0.12
 Unspecified substance use b 36 11 −0.23* −0.27, −0.18 36 11 −0.23* −0.27, −0.18
 Substance use consequences 9 5 −0.17 −0.40, 0.06 8 4 −0.09 −0.42, 0.24
Race/Ethnicity
 Hispanic 115 16 −0.21* −0.26, −0.16 109 16 −0.20* −0.25, −0.15
 Native American 65 7 −0.13* −0.23, −0.02 63 7 −0.12* −0.21, −0.03
 Black 41 7 −0.24* −0.37, −0.11 31 7 −0.20* −0.27, −0.13
Assessment/Follow-up timing
 0–12 months 153 20 −0.15* −0.22, −0.08 140 20 −0.15* −0.21, −0.08
 > 12 months 68 19 −0.22* −0.28, −0.17 63 19 −0.20* −0.23, −0.16
Open in a new tab

Note. K1 = number of effect sizes; k2 = number of studies; CI = confidence interval; outliers were removed in sensitivity analyses.

a

Illicit and other drug use includes cocaine, inhalant, prescription medication, methamphetamine, and other drugs.

b

Unspecified substance use refers to substance use composite variables/scores.

*

p < .05

Moderator Analysis

Meta-regression was conducted to explore whether the effects of culturally sensitive prevention programs for substance use outcomes varied across various moderators, including participant characteristics, intervention characteristics, type of substances, control conditions, assessment timing, and risk of bias. Results are presented in Table 3. Most of the meta-regression coefficients were trivial in magnitude and were statistically non-significant, with a few exceptions. The mean effect sizes were statistically significantly larger for cigarette smoking (b = −0.14, SE = 0.06) and unspecified substance use (b = −0.09, SE = 0.04) than for marijuana use. Culturally sensitive prevention programs with active control groups appeared to have larger mean effect sizes than those with non-active control groups (b = −0.10, SE =0.04). Regarding the risk of bias, higher risk of bias in the outcome measurement and selection of the reported results was associated with lower mean effect sizes (b = 0.12, SE = 0.04 and b = 0.06, SE = 0.02, respectively), holding other risk of bias domains constant. Given the expected number of included studies (i.e., 30), the average sample size per group (around 250), and moderate heterogeneity, I2 = 0.50, the power to detect two-group differences of 0.3, 0.2, and 0.1 in standardized mean differences based on random effects model is 99%, 86%, and 33%, respectively. Therefore, a moderator analysis would be underpowered to detect two-group differences less than 0.2 in Hedges’s g in the current study. Nevertheless, many of the non-significant effect size differences were small in magnitude (e.g., age, sex, duration, setting, addressed substance use versus positive youth development only), suggesting the differences are not meaningful even in the face of low statistical power.

Table 3.

Meta-Regression Results

Original Sensitivity
k1 k2 Coefficient SE k1 k2 Coefficient SE
Participant characteristics (Univariate)
Mean age (year) 221 30 −0.003 0.014 203 30 0.00 0.009
Sex (% female) 219 29 −0.002 0.002 201 29 −0.002 0.001
Ethnicity (reference group: Native American)a 221 30 203 30
  Hispanic −0.09 0.04 −0.08 0.04
  Black −0.11 0.06 −0.08 0.04
Intervention characteristics (Univariate)
Duration (hour) 218 29 0.002 0.002 200 29 0.002 0.002
Intervention type (non-parent based = 0, parent based = 1) 221 30 −0.07 0.04 203 30 −0.05 0.03
Intervention type (addressed substance use = 0, positive youth development only = 1) 221 30 0.005 0.06 203 30 −0.01 0.06
Intervention type (culturally adapted = 0, culturally grounded = 1) 221 30 0.06 0.04 203 30 0.05 0.03
Setting (reference group: School)b 221 30 203 30
  Community 0.004 0.05 −0.002 0.03
  Online −0.06 0.08 0.006 0.03
Control (non-active = 0, active = 1) 221 30 −0.10* 0.04 203 30 −0.08* 0.02
Assessment timing (0−12 months = 0, > 12 months = 1) 221 30 −0.06 0.04 203 30 −0.04 0.03
Outcome (reference group: Marijuana use)c 221 30 203 30
  Alcohol use −0.05 0.05 −0.01 0.04
  Smoking −0.14* 0.06 −0.09 0.06
  Illicit and other drug use −0.21 0.10 −0.11 0.06
  Unspecified substance use −0.09* 0.04 −0.06 0.04
 Substance use consequences −0.13 0.12 0.05 0.06
Risk of bias (Multivariate, low = 0, some concerns = 1, high = 2) 221 30 203 30
D1: Randomization −0.02 0.12 −0.01 0.11
D2: Deviations from intended interventions 0.04 0.03 0.02 0.02
D3: Missing outcome data −0.02 0.03 −0.01 0.02
D4: Outcome measurement 0.12* 0.04 0.08* 0.03
D5: Selection of the reported results 0.06* 0.02 0.04 0.02
Overall risk of bias 221 30 0.07 0.04 203 30 0.04 0.02
Open in a new tab

Note. k1 = number of effect sizes, k2 = number of studies, SE = Standard Error;

*

p < .05.

a

The mean effect size difference between Hispanic and Black adolescents was non-significant.

b

The mean effect size difference between community and online was non-significant.

c

The mean effect size difference between alcohol use and unspecified substance use was statistically significant under sensitivity analyses (b = −0.06, SE = 0.02).

Two potentially meaningful but non-significant effect size differences are the differences between racial groups (b = −0.09 and −0.11, p < .10, for Hispanic versus Native American, and Black versus Native American, respectively) and the differences between parent-based versus non-parent-based programs (b = −0.07, p < .10). Because many of the included programs for Hispanic adolescents (12 out of 16 included studies) and Black adolescents (4 out of 7 included studies) were parent-based or involved substantial parent participation and none of the included studies for Native American adolescents was parent-based, we further explored the racial group differences. After controlling for whether the intervention is parent-based, the racial group differences in effect sizes slightly decreased (b = −0.06 and −0.08, p > .10, for Hispanic versus Native American, and Black versus Native American, respectively). When looking at non-parent-based programs only, the effect size difference between Hispanic and Native American groups was trivial (b = −0.03, p > .10) and the difference between Black and Native American groups remained similar (b = −0.12, p > .10).

Sensitivity Analysis

Eighteen effect sizes were identified as outliers (i.e., when the individual effect size’s CI does not overlap with the CI of the pooled effect). After removing these outliers, the effect sizes in Table 2 were mostly comparable to those in the original analysis. For the moderator analysis results in Table 3, the effect size differences based on the control condition and risk of bias in outcome measurement remained comparable and statistically significant. The only statistically significant mean effect size difference based on the types of substance use outcomes was between alcohol use and unspecified substance use (b = −0.06, SE = 0.02).

Publication Bias

Figure 3 shows a funnel plot of all effect sizes in relation to the standard errors of the effect sizes. The trim-and-fill method estimated five effect sizes to be missing on the right side. The effect sizes were comparable between the observed effect sizes and the observed plus imputed effect sizes. However, due to the small number of included studies and the complicated nature of publication bias, we cannot conclude that our findings are robust against publication bias based on the funnel plot and the trim-and-fill method.

Figure 3.

Figure 3

Open in a new tab

Funnel Plot

Discussion

Summary of Intervention Effects

The present systematic review and meta-analysis of 30 RCTs found supportive evidence for the efficacy of culturally sensitive prevention programs (either substance use prevention or positive youth development) in preventing or reducing substance use among U.S. Black, Hispanic, and Native American adolescents. The average intervention effect size (g = −0.20) we observed was generally consistent with those found by previous meta-analyses for culturally sensitive substance use programs (Hernandez Robles et al., 2018; Hodge et al., 2012) but was more precise (i.e., having a narrower confidence interval) because we synthesized a greater number of effect sizes using robust variance estimation in meta-regression. Further, the small significant effects we found were comparable with but somewhat in between the effect sizes from previous meta-analyses of general school-based substance use prevention programs (Onrust et al., 2016) and family-based prevention programs for diverse adolescents (Van Ryzin et al., 2016).

The average treatment effects were statistically significant for all the substance use behavioral outcomes but not for substance use consequences. Very few included studies reported substance use problems or consequences as outcomes (i.e., nine effect sizes from five studies). Moderator analyses suggest that the average treatment effects appeared to be slightly smaller for marijuana use and alcohol use than for other types of substance use. These findings suggest that culturally sensitive prevention programs may have different magnitudes of effects for different types of substance use. These findings are somewhat consistent with previous meta-analysis results for school-based substance use prevention programs for early and middle adolescents (Onrust et al., 2016). However, another meta-analysis of family-based prevention programs found no significant differences in overall program effects across different types of substances (Van Ryzin et al., 2016). It is unclear what leads to this difference in our meta-analysis. Based on previous research, the differences might be attributed to the outcome targeted in the intervention (e.g., exclusively targeted cigarette smoking or targeted multiple substances or risk behaviors). Previous meta-analysis found that adolescent risk behaviors (including substance use) are modestly correlated (about 0.3), suggesting that targeting behavior-specific determinants may yield larger effects than only targeting common determinants of multiple risk behaviors (Guilamo-Ramos et al., 2005). Moreover, a previous meta-regression of school-based prevention programs found that while some program components (e.g., self-control, problem solving, healthy alternatives, parental involvement, social norm) were associated with both less smoking and alcohol use, some program components (e.g., social skills, peer education) were only associated with reduced smoking, and others only associated with reduced alcohol use (e.g., health education, refusal skills, cognitive behavior therapy, behavioral management; Onrust et al., 2016). Therefore, prevention programs that have various components targeting multiple risk and multiple substance use behaviors may yield somewhat different effects across types of substances. In addition, epidemiological data indicate that marijuana is increasingly displacing alcohol and cigarettes as the first substance use among adolescents who use multiple substances (Keyes et al., 2019), and alcohol remains the most commonly used substance among adolescents, which suggests that more prevention efforts are needed for adolescent marijuana use and alcohol use.

Subgroup analyses suggest that programs that were culturally tailored to Hispanic, Black, and Native American adolescents were effective in preventing or reducing their substance use. However, the average effect size was slightly lower (although not statistically significant) among programs tailored to Native American adolescents than programs for Hispanic or Black adolescents. It is unclear what leads to the effect size differences. One potential explanation is that many of the included studies for Hispanic and Black adolescents were parent-based or involve substantial parent participation, whereas none of the included studies for Native American adolescents was parent-based. When looking at non-parent-based programs, the effect size difference between Hispanic and Native American groups was much smaller. A previous review also highlights the lack of family-based prevention programs among Native American adolescents (Snijder et al., 2020). This previous review found that 9 out of the 18 substance use prevention programs for U.S. Native American adolescents had beneficial effects (Snijder et al., 2020). Moreover, literature has documented numerous challenges to developing culturally sensitive prevention programs for Native American people (King et al., 2009) and challenges to conducting RCTs among Native American communities (Dickerson et al., 2020). Future prevention programs for Native American adolescents should keep partnering with Native American adolescents and their communities to select the best intervention approach and research design collaboratively. Future systematic reviews may need to include both RCTs and studies using quasi-experimental designs to get a more comprehensive picture of the average program effects for Native American adolescents.

In addition, it is unclear why culturally sensitive prevention programs with active control groups appeared to have larger mean effect size than those with non-active control groups. This was unexpected because active control groups tend to control for part of the explanations for the intervention effects (e.g., some active ingredients of the intervention, attention from study staff, social support). The difference might be due to other design factors but we were unable to explore further due to the limited number of included studies that had active control groups. Because studies that had comparable active control groups were also rated as having a low risk of bias in the outcome measurement domain, the regression coefficients of control condition and risk of bias in outcome measurement were similar in magnitude. Other risk of bias domains appeared to have minimal influences on overall effect size in our study.

Summary of Culturally Sensitive Intervention Components

A growing number of prevention researchers have recognized the importance of integrating culturally specific content into prevention interventions to support participants’ understanding, identification, and receptiveness (Castro et al., 2004). The included studies’ descriptions of their processes of adapting existing programs or developing culturally grounded programs reflect the eight dimensions described in the ecological validity model (Bernal & Sáez-Santiago, 2006), which include language (e.g., offering intervention in Spanish and English for Hispanic adolescents, using culturally appropriate language), persons (e.g., bilingual and bicultural community group leaders), metaphors (e.g., indirect teaching using Native American stories, metaphors, cultural symbol of the medicine wheel), content (e.g., racial and cultural socialization, acculturation), concepts (e.g., Hispanic values of familism, respect, personal treatment, and sympathy), goals (e.g., increasing ethnic pride, learning adaptative behaviors to use when encountering racism), methods (e.g., using community-based participatory research, providing bilingual curriculum and materials, intergenerational community events, Spanish telenovelas), and context (e.g., materials relevant to inner-city settings, rural areas, immigration, structural barriers, and discrimination).

Overall Completeness, Applicability of Evidence, and Methodological Observations

The current review included studies conducted with urban Hispanic adolescents, urban and rural Black adolescents, and urban and rural Native American adolescents aged 11 to 18 years in the United States. Thus, the external validity of the review is limited to adolescents of these racial and ethnic groups and cannot be generalized to all adolescents of color. The reviewed studies compared various types and formats of culturally adapted or grounded prevention programs, covering individual-, parent-, school-, or community-based substance use prevention programs and positive youth development programs with no treatment, attention control, or alternative interventions. The results cannot be generalized to substance use treatments or non-culturally sensitive prevention programs. The reviewed studies covered self-reported adolescent substance use outcomes (i.e., ever used, frequency, amount, intentions, problems and consequences) and did not include toxicology screening outcomes.

The studies included in this review have methodological advantages as well as limitations. We examined the internal validity of the included studies. Overall, the included studies were judged to have either some concerns or have a high risk of bias based on RoB 2. The risk of bias may lead to either over- or under-estimation of the intervention effect sizes. To reduce these risks of bias, future clinical trials are encouraged to use comparable control groups, use centralized allocation concealment and describe their randomization process in detail, use outcome measures in addition to self-reported ones, register trial protocols with a detailed data analysis plan, increase the retention of participants, and properly use an intention-to-treat analysis. We did not formally assess the methodological qualities (e.g., external validity, precision) of the included studies but we have a few observations. First, most studies used nonprobability sampling, which limits generalizability beyond the studies’ sample. Second, a few studies had small sample sizes, which may not provide sufficient statistical power to detect small intervention effects. Third, about two thirds of the studies reported follow-up effects (i.e., more than 12 months post-treatment), which provide evidence of the long-term effects of the culturally sensitive prevention programs on substance use. Future clinical trials should employ probability sampling methods, recruit larger samples, and conduct follow-ups for longer periods of time when evaluating the efficacy of culturally sensitive prevention programs.

Strengths and Limitations of the Current Review

This review has several strengths as well as limitations. First, a comprehensive database and grey literature search were conducted to include relevant publications and intervention trials. Nevertheless, not all eligible studies may have been identified. Second, we conducted moderator analyses via meta-regression to explore whether the intervention effects differed by any study and participant characteristics. Such moderator analyses have been limited in previous reviews. However, the moderator analyses need to be interpreted with caution because (a) they are observational in nature and (b) non-significant coefficients cannot rule out the potential moderation effects due to a lack of statistical power (Hedges & Pigott, 2004). Third, the review used robust variance estimation to deal with dependencies among multiple effect sizes that were extracted from the same study, which allowed us to include a large number of effect size estimates. Fourth, we used the most updated risk of bias assessment tool (i.e., ROB 2) to assess the internal validity of the included studies and incorporated the risk of bias assessment results into data synthesis through meta-regression.

Among the limitations, we were unable to conclude the efficacy of culturally sensitive relative to non-culturally sensitive programs for substance use due to the limited number of included studies that had such control groups (i.e., active control group with the same key intervention components without cultural adaptation). Second, to preserve statistical power for the meta-regression, we aggregated effect size estimates based on type of substance rather than specific measures (e.g., alcohol use frequency, heavy episodic drinking, alcohol use intentions). When a large amount of effect sizes is available for each outcome, researchers are encouraged to focus on the more specific outcomes. Similarly, we were unable to test for more specific moderators or test moderators multivariately due to limited statistical power. Third, we only extracted effect size data from the published results, which likely are part of the results of these clinical trials (e.g., selected assessment timing, outcome measures). Future meta-analyses are encouraged to also obtain the original data.

Implications for Future Research

Most culturally sensitive prevention initiatives target many behavioral, developmental, contextual, and cultural factors related to substance use, making it challenging to identify the effective intervention components. For example, it is unclear if changes in cultural and developmental factors mediate the intervention effects on substance use behavioral outcomes. To keep a prevention program effective, cost efficient, and resource respectful, future clinical trials of culturally sensitive prevention programs should test the mechanisms of main intervention components so that unimportant factors can be dropped and active ingredients can be strengthened (Jaccard & Bo, 2018). In addition, because the meta-analyses findings suggest that the intervention effect may vary based on the type of substance and racial/ethnic group, it is necessary to measure and report the specific type of substance use outcome rather than mixing several types of substance use into one composite variable, and report intervention effects for the subgroup (e.g., sex, race and ethnicity, nativity status, substance use risk) of adolescents or examine moderators of the intervention effects. Finally, more research is needed to fill the substantial gap in developing and evaluating culturally sensitive prevention efforts for Asian American, Pacific Islander, and multiracial adolescents.

Conclusion

The current systematic review and meta-analysis synthesize the rapidly growing empirical evidence on culturally sensitive prevention programs for substance use outcomes among Hispanic, Black, and Native American adolescents. Findings indicate that culturally sensitive prevention programs that address substance-specific risks and protective factors, positive youth development factors, and reflect adolescents’ cultural values, norms, practices, and worldviews are beneficial for preventing or reducing substance use among Hispanic, Black, and Native American adolescents. More high-quality research is needed to replicate the findings and to advance the development and testing of culturally sensitive prevention interventions that are helpful for multiracial, Asian American, and Pacific Islander adolescents.

Supplementary Material

Appendix A. Characteristics of Included Studies
Appendix B. Included Studies
Appendix C. Risk of Bias Summary

Highlights.

  • Culturally sensitive programs can prevent substance use among Black, Hispanic, and Native American youth

  • Program effects may differ based on the type of substance

  • More studies are needed for Asian American, Pacific Islander, and multiracial youth

Role of funding sources:

Funding for this study was provided by National Institute on Drug Abuse (NIDA) Grant R01DA051578. NIDA had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication.

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.

Declarations

Ethics approval: The research was classified as non-human subject research by the Institutional Review Board from the authors’ institutions.

Conflicts of interest: None.

References

  1. Alegria M, Carson NJ, Goncalves M, & Keefe K (2011). Disparities in treatment for substance use disorders and co-occurring disorders for ethnic/racial minority youth. Journal of the American Academy of Child and Adolescent Psychiatry, 50(1), 22– 31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ayón C, Nieri T, & Ruano E (2020). Ethnic-racial socialization among Latinx families: A systematic review of the literature. Social Service Review, 94(4), 693–747. [Google Scholar]
  3. Bardo MT, Williams Y, Dwoskin LP, Moynahan SE, Perry IB, & Martin CA (2007). The sensation seeking trait and substance use: Research findings and clinical implications. Current Psychiatry Reviews, 3(1), 3–13. [Google Scholar]
  4. Benner AD, Wang Y, Shen Y, Boyle AE, Polk R, & Cheng YP (2018). Racial/ethnic discrimination and well-being during adolescence: A meta-analytic review. American Psychologist, 73, 855–883. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Belenko S, Sprott JB, & Petersen C (2004). Drug and alcohol involvement among minority and female juvenile offenders: Treatment and policy issues. Criminal Justice Policy Review, 15 (1), 3–36. [Google Scholar]
  6. Bernal G, & Sáez-Santiago E (2006). Culturally centered psychosocial interventions. Journal of Community Psychology, 34 (2), 121–132. [Google Scholar]
  7. Bo A, Hai AH, & Jaccard J (2018). Parent-based interventions on adolescent alcohol use outcomes: A systematic review and meta-analysis. Drug and Alcohol Dependence, 191, 98–109. [DOI] [PubMed] [Google Scholar]
  8. Bo A, Hai AH, Chen D, & Hammock K (2021). Risk of bias assessments in systematic reviews and meta-analyses of behavioral interventions for substance use outcomes. Journal of Clinical Epidemiology, 139, 20–27. [DOI] [PubMed] [Google Scholar]
  9. Bo A, Goings TC, Evans C, Sharma A, Jennings Z, & Sigmon G (2020). Culturally adapted substance use prevention interventions for racial and ethnic minority youth. National Institute for Health Research https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020215500
  10. Carbia CL, pez-Caneda E, Corral M, & Cadaveira F (2018). A systematic review of neuropsychological studies involving young binge drinkers. Neuroscience & Biobehavioral Reviews, 90, 332–349. [DOI] [PubMed] [Google Scholar]
  11. Castro FG, Barrera M, & Martinez CR (2004). The cultural adaptation of prevention interventions: Resolving tensions between fidelity and fit. Prevention Science, 5(1), 41–45. [DOI] [PubMed] [Google Scholar]
  12. Cave L, Cooper MN, Zubrick SR, & Shepherd CCJ (2020). Racial discrimination and child and adolescent health in longitudinal studies: A systematic review. Social Science & Medicine, 250, 112864. [DOI] [PubMed] [Google Scholar]
  13. Center for Behavioral Health Statistics and Quality. (2021). Racial/ethnic differences in substance use, substance use disorders, and substance use treatment utilization among people aged 12 or older (2015–2019) (Publication No. PEP21–07-01–001). Substance Abuse and Mental Health Services Administration https://www.samhsa.gov/data/
  14. Centers for Disease Control and Prevention (2019). Youth Risk Behavior Survey Data Available at: www.cdc.gov/YRBSS. Accessed on October 19th, 2022.
  15. Chartier K, & Caetano R (2010). Ethnicity and health disparities in alcohol related research. Alcohol Research & Health, 33(1–2), 152–160. [PMC free article] [PubMed] [Google Scholar]
  16. Covidence. (n.d.). Covidence Systematic Review Software Veritas Health Innovation. https://www.covidence.org
  17. Dickerson D, Baldwin JA, Belcourt A, Belone L, Gittelsohn J, Keawe’aimoku Kaholokula J, Lowe J, Patten CA, & Wallerstein N (2020). Encompassing cultural contexts within scientific research methodologies in the development of health promotion interventions. Prevention Science, 21(Suppl 1), 33–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Duval S, & Tweedie R (2000). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics, 56, 455–463. [DOI] [PubMed] [Google Scholar]
  19. Fisher S, Zapolkis TC, Sheehan C, & Barnes-Najor J (2017). Pathway of protection: Ethnic identity, self-esteem, and substance use among multiracial youth. Addictive Behaviors, 72, 27–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Grindal M, & Nieri T (2016). The relationship between ethnic-racial socialization and adolescent substance use: An examination of social learning as a causal mechanism. Journal of Ethnicity in Substance Abuse, 15(1), 3–24. [DOI] [PubMed] [Google Scholar]
  21. Guilamo-Ramos V., Litardo HA., & Jaccard J. (2005). Prevention programs for reducing adolescent problem behaviors: Implications of the co-occurrence of problem behaviors in adolescence. Journal of Adolescent Health, 36(1), 82–86. [DOI] [PubMed] [Google Scholar]
  22. Hawkins JD, Van Horn ML, & Arthur MW (2004). Community variation in risk and protective factors and substance use outcomes. Prevention Science, 5(4), 213–220. [DOI] [PubMed] [Google Scholar]
  23. Hedges L, & Olkin I (1985). Statistical methods for meta-analysis Academic Press. [Google Scholar]
  24. Hedges LV, & Pigott TD (2004). The power of statistical tests for moderators in meta-analysis. Psychological Methods, 9(4), 426–445. [DOI] [PubMed] [Google Scholar]
  25. Hedges LV, Tipton E, & Johnson MC (2010). Robust variance estimation in meta‐regression with dependent effect size estimates. Research Synthesis Methods, 1(1), 39–65. [DOI] [PubMed] [Google Scholar]
  26. Hernandez Robles E, Maynard BR, Salas-Wright CP, & Todic J (2018). Culturally adapted substance use interventions for Latino adolescents: A systematic review and meta-analysis. Research on Social Work Practice, 28(7), 789–801. [Google Scholar]
  27. Higgins JP, & Green S (Eds.). (2011). Cochrane handbook for systematic reviews of interventions version 5.1.0 [updated March 2011] https://www.handbook.cochrane.org [Google Scholar]
  28. Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, & Welch VA (Eds.). (2020). Cochrane handbook for systematic reviews of interventions version 6.1 [updated September 2020] https://www.training.cochrane.org/handbook. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hodge DR, Jackson KF, & Vaughn MG (2012). Culturally sensitive interventions and substance use: A meta-analytic review of outcomes among minority youths. Social Work Research, 36(1), 11–19. [Google Scholar]
  30. Hussong AM, Ennett ST, Cox MJ, & Haroon M (2017). A systematic review of the unique prospective association of negative affect symptoms and adolescent substance use controlling for externalizing symptoms. Psychology of Addictive Behaviors, 31(2), 137– 147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Jaccard J, & Bo A (2018). Prevention science and child/youth development: Randomized explanatory trials for integrating theory, method, and analysis in program evaluation. Journal of the Society for Social Work and Research, 9, 651–687. [Google Scholar]
  32. Keyes KM, Rutherford C, & Miech R (2019). Historical trends in the grade of onset and sequence of cigarette, alcohol, and marijuana use among adolescents from 1976–2016: Implications for “Gateway” patterns in adolescence. Drug and Alcohol Dependence, 194, 51–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. King M, Smith A, & Gracey M (2009). Indigenous health part 2: The underlying causes of the health gap. The Lancet, 374(9683), 76–85. [DOI] [PubMed] [Google Scholar]
  34. Landen M, Roeber J, Naimi T, Nielsen L, & Sewell M (2014). Alcohol-attributable mortality among American Indians and Alaska Natives in the United States, 1999– 2009. American Journal of Public Health, 104(S3), S343–S349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Leung RK, Toumbourou JW, & Hemphill SA (2014). The effect of peer influence and selection processes on adolescent alcohol use: A systematic review of longitudinal studies. Health Psychology Review, 8(4), 426–457. [DOI] [PubMed] [Google Scholar]
  36. Lewis B, Hoffman L, Garcia CC, & Nixon SJ (2018). Race and socioeconomic status in substance use progression and treatment entry. Journal of Ethnicity in Substance Abuse, 17(2), 150–166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Miech RA, Johnston LD, O’Malley PM, Bachman JG, Schulenberg JE, & Patrick ME (2021). Monitoring the Future national survey results on drug use, 1975–2020: Volume I, Secondary school students. Institute for Social Research, The University of Michigan http://monitoringthefuture.org/pubs.html#monographs
  38. Montes KS, Witkiewitz K, Pearson MR, & Leventhal AM (2019). Alcohol, tobacco, and marijuana expectancies as predictors of substance use initiation in adolescence: A longitudinal examination. Psychology of Addictive Behaviors, 33(1), 26–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Moore TH, Zammit S, Lingford-Hughes A, Barnes TR, Jones PB, Burke M, & Lewis G (2007). Cannabis use and risk of psychotic or affective mental health outcomes: A systematic review. The Lancet, 370(9584), 319–328. [DOI] [PubMed] [Google Scholar]
  40. Mulia N, Ye Y, Zemore SE, & Greenfield TK (2008). Social disadvantage, stress, and alcohol use among Black, Hispanic, and White Americans: Findings from the 2005 U.S. National Alcohol survey. Journal of Studies on Alcohol and Drugs, 69(6), 824–833. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Mulia N, Karriker-Jaffe KJ, Witbrodt J, Bond J, Williams E, & Zemore SE (2017). Racial/ethnic differences in 30-year trajectories of heavy drinking in a nationally representative US sample. Drug and Alcohol Dependence, 170, 133–141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Nair RL, Roche KM, & White RM (2018). Acculturation gap distress among Latino youth: Prospective links to family processes and youth depressive symptoms, alcohol use, and academic performance. Journal of Youth and Adolescence, 47(1), 105–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Neblett EW, Terzian M, & Harriott V (2010). From racial discrimination to substance use: The buffering effect of racial socialization. Child Development Perspectives, 4, 131–137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Nieri T, Ayón C, Yoo M, & Webb M (2020). Perceived ethnic discrimination, ethnic-racial socialization, and substance use among ethnic minority adolescents. Journal of Ethnicity in Substance Abuse, 21(1), 70–89. [DOI] [PubMed] [Google Scholar]
  45. Okamoto SK, Kulis S, Marsiglia FF, Holleran Steiker LK, & Dustman PA (2014). A continuum of approaches toward developing culturally focused prevention interventions: From adaptation to grounding. The Journal of Primary Prevention, 35, 103–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Onrust SA, Otten R, Lammers J, & Smit F (2016). School-based programmes to reduce and prevent substance use in different age groups: What works for whom? Systematic review and meta-regression analysis. Clinical Psychology Review, 44, 45–59. [DOI] [PubMed] [Google Scholar]
  47. Poe-Yamagata E, & Jones M (2000). And justice for some: Differential treatment of minority youth in the justice system. Building Blocks for Youth
  48. Sanchez D, Hamilton ER, Gilbert DJ, & Vandewater EA (2018). Examining Africentric cultural values, ethnic identity, and substance use abstinence in low-income, early adolescent, African American girls. Journal of Black Psychology, 44(1), 74–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Scheier LM, & Griffin KW (2021). Youth marijuana use: A review of causes and consequences. Current Opinion in Psychology, 38, 11–18. [DOI] [PubMed] [Google Scholar]
  50. Sirin SR, Choi E, & Sin EJ (2022). Meta-analysis on the relation between acculturation and alcohol use among immigrant youth. The Journal of Adolescent Health: Official Publication of the Society for Adolescent Medicine, 70(3), 361–377. [DOI] [PubMed] [Google Scholar]
  51. Snijder M, Stapinkis L, Lees B, Ward J, Conrod P, Mushquash C, Belone L, Champion K, Chapman C, Teesson M, & Newton N (2020). Preventing substance use among indigenous adolescent in the U.S.A., Canada, Australia and New Zealand: A systematic review of the literature. Prevention Science, 21, 65–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Sharmin S, Kypri K, Khanam M, Wadolowski M, Bruno R, Attia J, Holliday E, Palazzi K, & Mattick RP (2017). Effects of parental alcohol rules on risky drinking and related problems in adolescence: Systematic review and meta-analysis. Drug and Alcohol Dependence, 178, 243–256. [DOI] [PubMed] [Google Scholar]
  53. StataCorp. (2021). Stata Statistical Software: Release 17 StataCorp LLC. [Google Scholar]
  54. Sterne JA, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, Cates CJ, Cheng H-Y, Corbett MS, Eldridge SM, Emberson JR, Hernán MA, Hopewell S, Hróbjartsson A, Junqueria DR, Jüni P, Kirkham JJ, Lasserson T, Li T … & Higgins JP(2019). RoB 2: A revised tool for assessing risk of bias in randomised trials. BMJ, 366, 14898. [DOI] [PubMed] [Google Scholar]
  55. Tael‐Öeren M, Naughton F, & Sutton S (2019). The relationship between parental attitudes and children’s alcohol use: A systematic review and meta‐analysis. Addiction, 114(9), 1527–1546. [DOI] [PubMed] [Google Scholar]
  56. Tanner-Smith EE, & Tipton E (2014). Robust variance estimation with dependent effect sizes: Practical considerations including a software tutorial in Stata and SPSS. Research Synthesis Methods, 5(1), 13–30. [DOI] [PubMed] [Google Scholar]
  57. Unger JB, Ritt-Olson A, Soto DW, & Baezconde-Garbanati L (2009). Parent–child acculturation discrepancies as a risk factor for substance use among Hispanic adolescents in Southern California. Journal of Immigrant and Minority Health, 11(3), 149–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Unger JB, Sussman S Begay C, Moerner L, & Soto C (2020). Spirituality, ethnic identity, and substance use among American Indian/Alaska Native adolescents in California. Substance Use & Misuse, 55(7), 1194–1198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Van Ryzin MJ, Roseth CJ, Fosco GM, Lee YK, & Chen IC (2016). A component-centered meta-analysis of family-based prevention programs for adolescent substance use. Clinical Psychology Review, 45, 72–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Viechtbauer W, & Cheung MW (2010). Outlier and influence diagnostics for meta‐ analysis. Research Synthesis Methods, 1(2), 112–125. [DOI] [PubMed] [Google Scholar]
  61. Wallace JM (1999). The social ecology of addiction: Race, risk, and resilience. Pediatrics, 103, 1122–1127. [PubMed] [Google Scholar]
  62. Yap MB, Cheong TW, Zaravinos‐Tsakos F, Lubman DI, & Jorm AF (2017). Modifiable parenting factors associated with adolescent alcohol misuse: A systematic review and meta‐analysis of longitudinal studies. Addiction, 112(7), 1142–1162. [DOI] [PubMed] [Google Scholar]
  63. Zapolski TC, & Clifton RL (2019). Cultural socialization and alcohol use: The mediating role of alcohol expectancies among racial/ethnic minority youth. Addictive Behaviors Reports, 9, 100145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Zapolski T, Fisher S, Banks DE, Hensel DJ, & Barnes-Najor J (2017). Examining the protective effect of ethnic identity on drug attitudes and use among a diverse youth population. Journal of Youth and Adolescence, 46(8), 1702–1715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Zapolski TC, Pedersen SL, McCarthy DM, & Smith GT (2014). Less drinking, yet more problems: Understanding African American drinking and related problems. Psychological Bulletin, 140(1), 188–223. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix A. Characteristics of Included Studies
NIHMS1858800-supplement-Appendix_A__Characteristics_of_Included_Studies.docx (52.4KB, docx)
Appendix B. Included Studies
NIHMS1858800-supplement-Appendix_B__Included_Studies.docx (28.5KB, docx)
Appendix C. Risk of Bias Summary
NIHMS1858800-supplement-Appendix_C__Risk_of_Bias_Summary.docx (21.3KB, docx)

RESOURCES