'Keepin' It REAL in Mexico: An Adaptation and Multisite RCT Implementation of a Substance Abuse Prevention Curriculum for Adolescents in Three Mexico Cities.

The need for more effective prevention efforts in Mexico is clear and increasingly urgent, with important ramifications for the US. In the US immigrant adolescents from Mexico historically have had lower drug use rates than US-born Mexican Americans, but these trends are changing rapidly. To fill the gap in evidence-based universal prevention programs in Mexico, this study proposes to adapt, implement, and test the efficacy of kiR-a school-based substance use prevention program in use in the US and several other countries-among a sample of youth from the three largest urban areas in Mexico. This study will directly impact factors influencing sustainability - leadership, capacity, and relationship building. The research team is partnering with Dr. Medina-Mora and the National Psychiatric Institute of Mexico, which oversees substance use prevention nationwide. This study will build long-term relationships with Mexican scientists, educators, government agencies, social workers, and community stakeholders; demonstrate the feasibility of coordinated strategies for implementing school-based prevention programs involving national and local stakeholders; and will increase the capacity of schools and communities to deliver future interventions to additional students. keepin' it REAL (kiR), a SAMHSA recognized model program on the National Registry of Effective Programs and Practices (NREPP), is a school-based, manualized curriculum for substance use prevention among middle school students. The kiR curriculum includes: 10 weekly classroom lessons; 5 videos demonstrating drug resistance strategies that were scripted, acted and filmed by students; and school-wide booster activities (e.g., replaying the videos) to reinforce the classroom lessons. The curriculum is implemented by regular classroom teachers who receive a two-day long training. The curriculum includes the mix of elements found to have the largest effect sizes in school-based prevention interventions: those that teach comprehensive life skills to increase knowledge of substance use, address affective and motivational aspects of substance use (e.g., anti-drug attitudes, beliefs and values) and build risk assessment, decision making, and drug refusal skills. kiR emphasizes the mastery and expansion of a repertoire of the drug resistance strategies most commonly and effectively employed by adolescents: to refuse substance offers with a simple "no", to explain or provide an excuse to decline the offer; to avoid situations where substances might be offered; and to leave when substances are offered. kiR teaches a variety of communication techniques to reject risky or undesirable influences, and the flexibility to select appropriately from a repertoire of strategies in different situations. Youths equipped with multiple strategies can assess situations where substances are offered, decide which resistance strategies will be most effective, and try additional strategies (repertoire) if the first one does not work. There are several reasons the original kiR curriculum is an appropriate foundation for developing efficacious school-based prevention approaches in Mexico: its attention to Mexican American cultural and social contexts of substance use; evidence of the applicability of kiR's core components in Mexico; and its ideal developmental timing for implementation. There are also theoretical and empirical reasons that additional adaptation to the cultural and social contexts of Mexico is needed. The cultural communication patterns of youth and the context of substance use offers needs to be incorporated when adapting it for use in Mexico. Cross-sectional studies in Mexico show that kiR's core elements and approach are relevant and effective for Mexican youth from different regions. Surveys of Mexican high school youth found that kiR strategies were not only used in Monterrey and in Guanajuato State, but that the use of these strategies is associated with decreased alcohol and tobacco use. A longitudinal randomized pilot field trial of kiR in two Guadalajara middle schools reported significant effects in reducing use of alcohol, cigarette, and marijuana use, but in ways that differed for males and females, suggesting that additional tailoring of the curriculum to gendered substance use cultural and social contexts may be needed. In addition to testing the efficacy of kiR in Mexico, the potential moderating influence of gender role norms and violence will be explored. Despite rapid cultural changes in gender socialization in Mexico and women's attainment of higher levels of education and workplace participation, traditional gender role norms for women continue to be influential. Young women adhering to more traditional gender role norms also tend to use substances less because females are more closely monitored by parents, more restricted to the home, and less exposed to risky environments such as unsupervised parties or bars. However, among the growing proportion of female adolescents in Mexico who are questioning traditional gender role norms, these cultural restraints are lessening and the risk of substance use is growing. Female and male adolescents in Mexico are nearing parity in rates of alcohol use, although females still report less use. Norms for men that promote hyper-masculinity, aggressiveness and impulsiveness encourage behaviors like heavy drinking and risky sex that have severe health consequences. The persistence of traditional male gender role norms underscores the pressing need for effective drug resistance strategies that provide male adolescents with ways to counterbalance unhealthy cultural messages with positive cultural norms. Gender differences among adolescents in Mexico have been reported in the use of the REAL drug resistance strategies, in the association between use of these strategies and actual substance use, and in the effects of kiR in a pilot feasibility field trial in Guadalajara. While both males and females used the REAL strategies to resist substance use offers, the relationship between substance use and the use of REAL strategies was either stronger for males than for females or only significant for males. The field trial found short-term effects in reducing females' use of alcohol and cigarettes and in the long-term use of alcohol, while desirable effects for males were found for long-term use of marijuana. The pervasive violence in some Mexican communities has the potential to influence youths' vulnerability to using substances and the effectiveness of strategies to prevent or resist substance use. Experiencing violence as a perpetrator, victim, or witness is a complex and interconnected phenomenon. Among urban minority youth in the US, 95% of perpetrators also report violence victimization or witnessing community violence. Although an investigation of the mechanisms linking violence to youth substance use is beyond its scope, the proposed study will investigate how youths' perpetration, victimization, and witnessing of violence may moderate the efficacy of kiR. This is a highly innovative study which will conduct the first national RCT of a school-based and evidence-based universal prevention intervention that is systematically culturally adapted for Mexico, using a NREPP model program. The proposed cultural adaptation design is innovative because it includes a bi-national team of stakeholders: students, teachers, school administrators, and researchers participating on an equal footing throughout the adaptation and testing phases. Second, the study advances knowledge on the transferability to emerging economies of prevention interventions that were developed, implemented, and tested in the US and based on a science of prevention. Third, the study will investigate and incorporate the gendered nature of substance use risk in the design of an adapted universal prevention program, and assess the differential efficacy of the program on male and female adolescents who adopt more and less traditional gender role norms in a societal context where those norms are rapidly changing. Fourth, the study will examine if Mexican adolescents experiencing different types of violence (exposure, perpetration, and victimization) derive differential benefits from universal school based substance use prevention programs like kiR. Fifth, a bi-national research team with a long-standing history of collaboration and productivity will lead the study, bringing together leading scientists in prevention in Mexico with the original designers of the kiR intervention. The ASU-based research team has a long trajectory starting in the 1990s related to the original development, testing, sub-group analysis, and dissemination of kiR. Over more than a decade, the research team has conducted three exploratory studies across different regions of Mexico in order to advance understanding of the social and cultural contexts of substance use among Mexican youth and the bi-national applicability of kiR. Study 1. A cross-sectional study conducted in Monterrey, Mexico with 327 students in two high schools (grades 10-12) demonstrated: the use of the REAL strategies to resist alcohol, tobacco and marijuana offers; gender differences in use of the strategies; and an inverse relationship between use of the strategies and actual substance use. Study 2. Similar findings emerged in a larger probability-based sample conducted in Guanajuato (State) with 702 students in eight high schools. The Monterrey and Guanajuato studies provided initial evidence that the use of REAL strategies is associated with less substance use among Mexican youth; however, the studies did not directly address the effectiveness of kiR in a Mexican cultural context. Study 3. To test the cultural fit of kiR in Mexico and to prepare for possible adaptation of the program to the Mexican context, a longitudinal randomized pilot feasibility trial was conducted in Guadalajara, Mexico with 432 middle school students in two schools. In this study, kiR was implemented with only minor linguistic adaptations (e.g. English-to-Spanish translation, Spanish dubbing of videos) and showed promising signs of effectiveness. The intervention had significant short-term effects in reducing females' use of alcohol and cigarettes, and the long-term use of alcohol, and desirable effects for males were found for long-term use of marijuana. The proposed study design is guided by communication competency theory and aims at testing the main hypothesis that by incorporating elements of the social and cultural contexts in which children live, prevention interventions will be more efficacious. Phase 1 (2 years): Aim 1, in this adaptation phase, the original kiR will be implemented in one school each in Guadalajara, Mexico City, and Monterrey. Students will participate in guided discussions and focus groups assessing the applicability in the Mexican context of the curriculum lessons' examples, substance offer scenarios, and drug resistance strategies. Implementing teachers from all sites will also complete feedback forms after each lesson and meet later in focus groups to discuss the implementation of the curriculum and suggest revisions. Student and teacher focus groups will be transcribed and analyzed to identify recurring themes and recommendations. Project Coordinators from each site and developers of the original curriculum will meet to decide on and incorporate revisions of the original curriculum to ensure cultural relevance. Special attention will be given to scenarios related to gender differences in drug offers and appropriate responses to them. Phase 2 (3 years): Aims 2, 3 \& 4, a longitudinal RCT of the adapted version of kiR will be conducted in 36 public middle schools in Guadalajara, Mexico City and Monterrey, with schools randomized within each city into three conditions: (A) culturally adapted kiR, (B) original kiR, and (C) control. In addition to questions relating to substance use and its psychosocial antecedents, pretest (T1), post-test (T2) and long-term follow-up surveys (T3) will assess traditional gender role norms and violence (victimization, perpetration and witnessing). The curriculum will be modified through a targeted adaptation process in which the original kiR is implemented in translated form in one school in each study site. Prior to program initiation, implementing teachers will receive training provided by the site Project Coordinators, the Mexican co-PI and supported by ASU master trainers. After each weekly lesson of the curriculum, students will complete a short curriculum feedback form and then participate in group discussions facilitated by a project team member and recorded. This will allow the research team to identify and address cultural and practical barriers to implementation and areas that need to be adapted or changed. State and local educational authorities as well as school principals will be invited to participate in the teachers' focus groups to identify system level issues with implementation, dissemination and sustainability. After curriculum feedback data have been collected, transcribed, and analyzed, Project Coordinators from each site will meet with members of the original US curriculum development team to incorporate the examples and scenarios provided by students. Core components of the curriculum will be preserved, e.g. overall structure, topics, and skills training in the use of the REAL strategies. During the curriculum adaptation process, the study's Bi-National Scientific Leadership Team led by the Mexican co-PI will review and approve suggested changes to ensure the cultural applicability and preservation of core components of kiR. After the adaptation is complete, the randomized control trial will be conducted in Guadalajara, Mexico City, and Monterrey, with 12 schools from each site assigned at random into three conditions: (A) culturally adapted kiR, (B) original kiR (translated only), and (C) control. All students will complete a pretest survey before the start of the intervention (T1) and post-tests to assess short-term (T2) and long term (T3) program effects on substance use behaviors and antecedents (drug use intentions, norms and attitudes), and drug resistance strategies. Sites, Schools, and Participants. The Mexico City, Monterrey, and Guadalajara metropolitan areas will serve as the three sites for the study. These sites were chosen as the largest metropolitan and economic hubs in Mexico (accounting for over a fourth of the Mexican population), as major sending communities to the US, and as prime destinations or stop-overs for migrants coming from rural Mexico, Central and South America, many on an intended migration path to the US. From these three sites, the sample of 7th grade children will be drawn from 36 public schools (Total N ≈ 7,768). The nine schools within each site will be randomized into conditions A, B and C, three schools per condition. Implementation in each site will be led by the research partners in México: 1) Instituto Nacional Psiquiatrico de México - Ciudad de México, 2) Universidad Autónoma de Nuevo León, and 3) Instituto Milton Erickson Guadalajara. Participant Recruitment. At the beginning of the school year, schools will distribute a general "letter-to-parents" that describes the study. About a week later, and with the schools' assistance, a letter to the 7th grade parents will be sent home via the children. This letter will again describe the study but will also invite parents to indicate if the child will not participate. Children whose parents do not want the children to participate will be excluded from all surveys. Children whose parents do not opt out will be randomized according to the treatment condition assigned to the child's school. Data Collection. Data collection will be the responsibility of the Mexican research partners, assisted by data collection, instrumentation, and quality assurance protocols developed by the US research team. A baseline pretest survey (T1) will be administered to consented 7th grade students at all sites at the start of the school year and before the beginning of kiR. The 10-lesson intervention will begin immediately after. An initial post-test (T2) will be administered two months after implementation ends (the end of 7th grade), at a time when the program effects on short-term psychosocial antecedents of substance use (e.g., drug-related intentions, norms,and attitudes) are expected to be strongest. An additional follow-up posttest (T3) will be administered one year later, at the end of 8th grade, to assess the long-term efficacy of the intervention. Fidelity measures will be collected during implementation. Project team members will observe implementation on a standardized schedule to ensure fidelity to kiR program. Participant Database and Tracking. The Mexican research partners will enter the data at each site and securely transfer it to the US research team. The US team will develop a database using REDCap which will track youth throughout the RCT. Unique identifiers will allow for matching across survey waves. The database will contain the participant's name, unique identifier, and site, school and teacher when first entering the study. Participant tracking through REDCap has been used successfully in previous studies by the team. Process Evaluation Measures. In the adaptation phase, teacher and student written feedback after each lesson will be self-reported on lesson feedback forms, using instruments and measures initially developed for the original kiR trial. In the RCT phase, the process of recruitment of the schools in the three cities will be documented to capture challenges and solutions. Fidelity is captured by items assessing the implementer's adherence to the program manual when delivering the lesson's key components-instructions, video, practice, and homework-as well as the implementer's enthusiasm, preparedness, implementation self-efficacy, and program acceptance (implementer comfort using program). Survey Measures. Key measures in the survey instrument are connected to communication competence theory to assess the repertoire of drug resistance strategies that the students use in different social and cultural contexts and how each may influence substance use. Traditional measures of substance use behaviors and the psychosocial antecedents were chosen for internal consistency over time through prior evaluations of kiR and use in other prevention trials, and are supplemented by measures of gender role norms and experience of violence. Table 1 summarizes key measures and psychometric properties. Spanish language back-translated versions of most items were employed successfully in the research team's preliminary pilot Mexico studies in Monterrey, Guanajuato, and Guadalajara (see Section C1). Prior to employing the questionnaire items in the proposed RCT, the items will be tested during the Phase I pilot test in three schools, one in each city, to reconfirm suitability, accurate translation, and psychometric properties. Substance Use Behaviors: Key student outcomes include the youths' self-reports of recent (last 30 days) frequency and amount of substance use (alcohol, cigarettes, marijuana, and inhalants). While no independent validation of actual substance use will be obtained through biochemical measures, there is sufficient evidence of the validity of self-reports under the proposed conditions: comparing self-reports over time and across treatments, focusing on recent substance use, and using administrative procedures that minimize bias (e.g., establishing rapport, insuring confidentiality, placing sensitive questions toward the end). Substance Use Antecedents: Key psychosocial antecedents of substance use will be assessed as program outcomes to test whether the kiR intervention reinforces important protective factors against substance use: substance use intentions, attitudes (positive expectancies of substance use), antidrug norms (personal, injunctive, and descriptive), and drug offer refusal confidence (or efficacy). A 3-item measure will assess the students' intentions to accept offers of alcohol, cigarettes or marijuana if offered to them the coming weekend. The positive substance use expectancies measure assesses perceptions of the positive consequences of substance use, e.g. whether using substances makes it easier to be part of a group, reduces nervousness, or makes parties more fun. Substance use personal norms will be gauged by asking: "Is it OK for someone your age to..." "drink alcohol," "smoke cigarettes," or "smoke marijuana". The injunctive drug norms items measure expected negative or positive reactions by parents and best friends to the respondents' use of substances. The items ask how angry the students' parent(s) would be if they found out that the student "smoked marijuana," "smoked cigarettes," or "drank alcohol", and how best friends would act towards them if they used the same substances. Two descriptive norms items (perceptions of the extent of substance use among age peers) ask respondents to estimate the proportion of school peers, and of friends, who have tried alcohol, cigarettes, and marijuana. Refusal confidence is measured by three items that ask the extent to which the student is sure that he or she would decline an offer of alcohol from a family member, of marijuana from a close friend, and of a cigarette from a school peer. Drug Resistance Strategies: The drug resistance strategies measures will assess students' responses to actual and hypothetical substance use offers: the frequency or likelihood of turning down an offer by "just saying no" (refuse); give an explanation or excuse (explain); stay away from situations where people the same age are using substances (avoid); leave the situation (leave); or some other resistance strategy. These items will be posed in two ways: as questions about resistance strategies actually used in the last 30 days, and as likely responses in a hypothetical situation, being offered a beer at a party next weekend. Gender Role Norms: Adherence to traditional gender role norms will be measured by the five-item Gender Roles subscale of the Mexican American Cultural Values Scales for Adolescents. The scale assesses adolescents' beliefs in having distinct roles for males (bread-winner and head of household) and females (child-rearing and protecting daughters). The subscale was validated and shown reliable with a sample of 750 fifth grade Mexican-heritage males and females in the Southwestern US. Exposure to Violence: A modified version of the Exposure to Violence and Violent Behaviors Committed checklists will be used to measure middle-school students' violence perpetration and violence victimization in the home, school, and neighborhood context. A modified 10-item subscale of Things I Have Seen and Heard will assess exposure to community violence. Demographics and Controls: Finally, a set of socio-demographic and control questions will include gender, usual academic grades, age and birth date (the latter useful for verifying tracking and matching across waves), and family and household composition. Socioeconomic status will be measured by utilizing information about the educational attainments of the respondents' parents and reports of household financial strain. Process Evaluation Analysis. The goal of the process evaluation in the adaptation phase is to identify further adaptations that are necessary to refine kiR for use in Mexico. The process will involve assembling, extracting, coding and comparing information from the open-ended discussion group and focus group data, and lesson feedback provided by students and teachers. Investigators will follow strategies that are designed to increase the rigorousness of analysis and interpretations including triangulation, analyst cross-checking, and an audit trail. Overall assessments of the salience of these patterns will then inform decisions about cultural adaptations to kiR that should be incorporated into the revised Mexican version of the curriculum. RCT Program Efficacy Analyses. Student questionnaire data will be employed to test hypotheses about the efficacy of the adapted kiR curriculum, using intent-to-treat analyses comparing all students at pretest who were in condition A schools with (1) students in condition B schools, and with (2) students in condition C schools. Prior to conducting hypothesis tests, measures will be assessed for adequate variability and scale reliability, and exploratory factor analysis will be employed in data reduction. For the hypothesis tests, the dependent variables will include T2 and T3 post-test measures of the outcomes listed above in section C5. Hypothesis 1. Program efficacy: students receiving the culturally adapted version of kiR will report relatively lower undesirable outcomes (e.g. recent substance use, intentions to use substances) and relative increases in desirable outcomes (e.g. anti-drug personal norms and attitudes, use of REAL strategies) when compared to students receiving the original kiR (translated only) and those in the control group. Hypotheses 2a, b. Gender role norms as moderators: The efficacy of the intervention will vary for males and females who differ in adherence to traditional gender role norms. Hypothesis 3. Violence as a moderator: The intervention will produce stronger desired effects among youth who report more experience of violence (as victim, perpetrator, or witness). These youth may benefit more from the intervention because of being at higher risk for using substances. Investigators will use both baseline adjusted general linear model procedures (i.e., variants of multiple regression) and growth curve modeling to test Hypothesis 1 concerning program efficacy. The first method will predict short-term (T2) and then long-term (T3) post-test outcome measures (substance use, antecedents, and drug resistance strategies) with dummy variable contrasts of intervention conditions, while controlling simultaneously for the baseline pre-test (T1) measure of the particular outcome. Condition A (adapted kiR) will be compared with Condition B (original kiR) and then compared in a separate analysis with Condition C (Control). Growth curve models will test whether and how changes over time in the outcomes (trajectories from T1 to T2 to T3) differ significantly by these same contrasts of conditions. Hypotheses 2a, 2b and 3 will be tested employing mean centered interactions of the measures of traditional gender role norms and experience of violence with the dummy variable contrasts of study conditions (A v. B, A v. C). Prior to these tests of moderation investigators will check the applicability of the assumption under the risk moderation argument that substance use behaviors (measured before intervention) vary according to traditional gender role norms and experience of violence as described in sections A5 and A6. The data will be analyzed using multilevel statistical models (e.g., SAS Proc Mixed), with individual students clustered within schools, to adjust for any school-level random effects that may result in biased standard errors. All analyses will introduce controls for the metropolitan site of the intervention and any demographic characteristics of participants that are shown to differ significantly across sites at the pretest. Missing Data, Attrition and Retention Strategies. As in any longitudinal study, there is likely to be attrition and item missing data, but our preliminary research suggests that it will be readily manageable with available statistical techniques. Investigators will employ multiple imputation methods to adjust for missing data, which have been used successfully in studies of prevention program efficacy. Using the PROC MI procedure in SAS, creating multiple complete datasets, analyze them with complete-data methods, and combine the results of these analyses to arrive at a single estimate that properly incorporates the uncertainty in the imputed values. The number of imputations will be adjusted, as recommended. Statistical Power. The sample size of 7,768 students within 36 schools was determined through a power calculation to be able to detect small to medium effect sizes. A minimal expected effect size of 0.15 for main outcomes was determined through the Guadalajara pilot feasibility study and prior US-based studies testing adaptations of kiR. Interclass correlations (ICCs) were calculated using data from the original RCT of kiR in 35 schools in the Southwestern US (Note: with only two schools, ICCs could not be calculated in the Guadalajara pilot feasibility study). Among the twelve main outcomes, the highest ICC was 0.02. Thus, given an ICC of 0.02, an effect size of 0.15, and a moderately effective pretest that explains 60 percent of the between school variation, our design has 0.80 power with 12 schools per treatment, 150 students per school (n = 7,768). Power was calculated for a two-level cluster randomized design using RDPOWER. Over the last decade the US research team has systematically developed research partnerships with colleagues in different regions of Mexico to better understand the substance use needs of Mexican youth and to test the applicability of kiR in the Mexican cultural and social context. These partnerships have also enriched the team's work in the US with Mexican American youth. The proposed study is a natural outgrowth of these bi-national collaborations. Mexican researchers who have collaborated with the US team on prior pilot studies in Mexico will also participate in the proposed study, advancing the partnership and enhancing prevention research capacity in Mexico. Responsibilities for completing the research will be shared by the Mexican and US research teams. Dr. Medina-Mora will serve as the Mexican Co-PI and will oversee the quality assurance procedures including fidelity and research protocols. Dr. Medina-Mora is a national and international distinguished researcher and Director of the National Psychiatric Institute of Mexico. The three teams of Mexican researchers will be responsible for recruitment and retention, collection of process evaluation and student survey data, and data entry at each respective site. The US research team will design the protocols for data collection, instrumentation, and quality assurance, and conduct the statistical analyses of program efficacy and tests of hypotheses. Both the US and Mexican teams will be involved in the curriculum adaptation process, interpretation of findings, manuscript development and dissemination activities.