Risk and Protective Factors for Adolescent Substance Use - Journal of ...

Journal of Adolescent Health 40 (2007) 448 – 455

Original article

Risk and Protective Factors for Adolescent Substance Use: Findings from a Study in Selected Central American Countries Wendy Kliewer, Ph.D.a,* and Lenn Murrelle, M.S.P.H., Ph.D.b b

a Department of Psychology, Virginia Commonwealth University, Richmond, Virginia Department of Epidemiology and Community Health, Virginia Commonwealth University, Richmond, Virginia Manuscript received July 5, 2006; manuscript accepted November 22, 2006

Abstract

Purpose: To identify the prevalence of substance use and problems with use, and risk and protective factors at different levels of the adolescent’s ecology associated with substance use among adolescents in selected Central American countries. Methods: Results of a survey of 17,215 students from Panama, Costa Rica, and Guatemala conducted in 2000 –2001 served as the basis for the analyses. Lifetime use of alcohol, tobacco, marijuana, and five other drugs (inhalants, tranquilizers, cocaine, crack, and ecstasy), and problems with drugs and alcohol were the outcome variables. Risk factors included dysregulation, family problems with drugs/alcohol, negative family interactions, school disengagement, peer deviance, and exposure to community violence. Protective factors included a personal belief in God, positive family interactions, parent religiosity, and positive student-teacher interaction. Both hierarchical linear regression and logistic regression analyses were used to model main and interaction effects of risk and protective factors. Results: There was a linear association between number of risk and protective factors and substance use, however, risk factors were more strongly associated with substance use than were protective factors. There were significant risk-by-protective-factor interactions for alcohol and marijuana use, and for problems with drugs and alcohol. Risk interacted most consistently with a personal belief in God, but also with parent religiosity and with student-teacher communication. Conclusions: It is important to consider risk and protective factors at different levels of an adolescent’s ecology. Prevention and intervention efforts should focus on interactions adolescents have in different microsystems (e.g., with parents, teachers, and peers). © 2007 Society for Adolescent Medicine. All rights reserved.

Keywords:

Risk; Protection; Central America; Substance use; Adolescents

Recent changes in Central America, including an increase in drug trafficking and guerilla violence, have raised concerns about drug use and abuse among the adolescents in those regions. Past research among Central American adolescents from the PACARDO Project has shown that among adolescents aged 12–20 years who were attending school, over half (51.5%) had tried alcohol, 29.1% had tried tobacco, and 4.6% had tried marijuana [1]. (The PACARDO *Address correspondence to: Dr. Wendy Kliewer, Department of Psychology, Virginia Commonwealth University, PO Box 842018, Richmond, VA, 23284-2018. E-mail address: [email protected]

name concatenates PA for Panamá, CA for Centroamérica, and RDO for República Dominicana. PACARDO refers to the name of the questionnaire used to assess drug use among high school students in those regions of the world.) Our approach in the present study was to evaluate both substance use, including alcohol, tobacco, marijuana and hard drug use, and problems associated with use. Problems associated with use, rather than frequency of use per se, aid in identifying individuals with substance use disorders (SUDs). We assessed a range of risk and protective factors associated with these outcomes. Based on the socio-ecological framework of Bronfenbrenner [2], we chose to examine risk and protective factors

1054-139X/07/$ – see front matter © 2007 Society for Adolescent Medicine. All rights reserved. doi:10.1016/j.jadohealth.2006.11.148

W. Kliewer and L. Murrelle / Journal of Adolescent Health 40 (2007) 448 – 455

in different microsystems. Microsystems are those contexts in which individuals have face-to-face interactions with others. Common microsystems associated with adolescents include family, school, peer, and neighborhood. The interactions in one microsystem affect interactions in other microsystems; Bronfenbrenner referred to this as a mesosystem. Understanding risk and protective factors in microsystems complement qualities of risk or protection individuals bring by virtue of temperament or other dispositional qualities. Because of the importance of both risk and protective factors in predicting adolescent substance use [3–5], we assessed both types of factors using multi-dimensional items. The risk factors selected for this study included dysregulation, family problems with drugs and/or alcohol, negative family interaction patterns, school disengagement, and peer deviance. Past research has identified each of these constructs as risk factors for adolescent problems with drugs and alcohol. In the past several years there has been increased attention to the construct of dysregulation. Dysregulation refers to a deficit in one’s ability to modulate affect, behavior, and cognition in response to environmental context resulting in ineffective execution of goal-directed plans [6]. For example, adolescents who have problems with dysregulation often are impulsive, and make poor choices without considering the consequences of their behavior. There is a vast literature on family problems with drugs and/or alcohol as a risk factor for adolescent substance use [7,8]. Parents or other family members who have problems with drugs and alcohol may model drug use, enhance opportunities for drug or alcohol use by having substances available, or fail to monitor adolescents’ behavior, thus increasing opportunity for drug and alcohol use. Likewise, negative family interaction patterns, particularly conflict and poor communication, have been linked to adolescent substance use problems [7–9]. Adolescents in conflicted families may not have the support to deal with stressors in their lives, or may experience additional stress as a result of conflictual family interaction, and thus may use drugs or alcohol as a form of stress relief. Further, families with poor communication may not be effective at monitoring adolescents’ behavior. School disengagement and peer deviance both contribute to and are consequences of adolescent substance involvement [7,10]. School disengagement typically starts well before mid-adolescence, and often co-occurs with learning difficulties. Once disengaged from school, adolescents may seek affiliation with deviant peers. It is in this context that most adolescent substance involvement begins. Finally, exposure to community violence places adolescents at risk for substance use involvement by virtue of the stress associated with being victimized by and witnessing community violence, and by opportunities for drug use that often occur in highviolence communities [11,12].

449

The protective factors included in this study are a personal belief in God, positive family interaction patterns, parent religiosity, and positive student-teacher interaction. Each of these constructs has been associated with positive adjustment during adolescence. Several large survey studies consistently have shown that religious adolescents engage in less deviant behavior, including substance use, relative to less religious adolescents [13]. This may be a function of personal beliefs regarding behavior, or constraints on behavior or support for healthy behavior from religious institutions. Likewise, parents who are religious tend to have adolescents who are religious [13,14], and those parents communicate values regarding behavior to their offspring. Further, religious parents are more likely to both support and monitor their adolescent children [14], thus decreasing both the need for substance use involvement and the opportunities to engage in substance use. Positive family interaction, specifically family cohesion and communication, is associated with a range of adaptive behaviors during adolescence, including lower levels of substance use involvement [15–17]. Cohesive families provide support for adolescents, and a context in which to learn, enact, and be reinforced for adaptive coping behavior. Work by Renick et al [17] using data from the National Longitudinal Study of Adolescent Health (Add Health) has shown that feeling cared for and connected to parents is highly protective against adolescent substance use. Finally, teachers can be a tremendous source of support for adolescents [17,18], particularly when adolescents lack support from parents. Teachers can provide a nonjudgmental point of view, can help adolescents feel connected to school, and can buffer negative interactions with peers. Resnick and colleagues [17] have shown that connection to school was an important protective factor for a range of adolescent outcomes, including substance use. An additional facet of the study involved examining not only the unique contributions of risk and protective factors to adolescent substance use, but the interactions among risk and protective factors. Recent discourse on resilience [19] has identified different patterns of protection associated with resiliency. In addition to main effects of protective factors (i.e., similar ameliorative effects across low- and high-risk conditions), protective factors may interact with risk status to produce a stabilizing effect (stability is conferred despite increasing risk), an enhancing effect (adolescents engage with the stressor such that competence is augmented with increasing risk), or a reactive effect (the protective factor generally confers advantages but less so when risk levels are high vs. low). The present analyses explore risk and protection associated with adolescent substance use and problems with use with continuous-level data as well as by constructing risk and protective factor indexes from counts of the number of risk and protective factors present.

450


Methods Study Population Data were obtained from a survey of Central American adolescents conducted in 2000 –2001 [20]. A total of 17,215 young people aged 12–20 years (M ⫽ 15.4 years, SD ⫽ 1.7 years; 51.6% female) from Panama (n ⫽ 4757), Costa Rica (n ⫽ 4948), and Guatemala (n ⫽ 7505) and participated. Two-thirds (66.4%) of the sample lived with both biological parents, which is consistent with data from other Central and South American samples. Additional data from highrisk samples also were obtained in the study, but are not considered here. A stratified random sampling methodology with some modifications was employed. In an effort to draw study samples that were broadly extrapolatable to their respective country, the most populated regions of each country were identified and used as the sampling frame. For each country, the universe of students from which samples were taken was representative of at least 85% of all students matriculated nationally. Ultimately, the provinces of San José, Heredia, Puntarenas, and Limón were selected for inclusion in Costa Rica , whereas the provinces of Panamá, Colón, Chiruqui, and Bocas del Toro were selected in the Republic of Panamá. In Costa Rica, a small number of schools in regions satisfying sampling criteria were eliminated from consideration as a result of recent student participation in another intensive survey study on drug use. In Guatemala, the departments of Alta Verapaz, Baja Verapaz, Chimaltenango, Chiquimula, El Progreso, Escuintla, Guatemala City, Huehuetenango, Izabal, Jalapa, Jutiapa, Peten, Quetzaltenango, Quiche, Retalhuleu, Sacatepequez, San Marcos, Santa Rosa, Solola, Suchitepequez, Totonicapan, and Zacapa were selected. Schools and classrooms from the above provinces were then randomly selected using multistage cluster sampling stratified by age, gender, and geographic region. All IRB standards in these three countries were met. Passive consent was used, consistent with the protocols used by the Ministries of Education in each country. Two weeks before the day of the study, parents received a letter from the school principal explaining the study and giving them the opportunity to “opt out” on behalf of their children. Parents who did not wish their child to participate returned the consent form to the school with their disapproval. Students also had the opportunity to opt out of the study on the day of testing. All students present in the selected classrooms on the day of the survey who had not opted out of the study were included in the target sample. Less than 1% of the students chose not to participate, which is consistent with other epidemiological research in Latin America. Absenteeism also was low (⬃2–3%) and students who were absent on the day of the survey were not replaced. Research assistants, most with professional level education and who were specifically trained in the goals and

methods of the study, informed the students carefully about the study and gave students an opportunity to refuse or to discontinue participation at any time. No compensation was offered for participation in the study. Research assistant training was manualized to provide consistency within and across countries. Research staff was available to answer the questions of individual students. Official school personnel, including classroom teachers, were absent during the average 40-minute administration. Research assistants underscored that the students’ responses were confidential, meaning their responses would not be seen by anyone other than research staff. Students were not allowed to write their names on the questionnaires and were cautioned not to look at the responses of their peers. These conditions are known to promote valid responses by adolescents to drug-abuserelated questionnaires [21]. Cooperation was greater than 95% in all three countries. Measurement The survey instrument consisted of approximately 175 multiple-choice questions (there were slightly variations across countries). Domains assessed in the survey included: alcohol, tobacco and other drug use, substance use problems, gang involvement, engagement in violence, depressive symptoms, dysregulation, family problems with drugs/ alcohol, religiosity, negative and positive family interactions and communication, school disengagement, positive studentteacher interaction, support from adults, peer deviance, and exposure to community violence. The final survey instrument used was constructed after a pilot study with a Panamanian sample of 988 adolescents aged 11–19 years [22]. Most of the instruments or items used in the current study had been used extensively in epidemiologic drug abuse research in Latin America and the Caribbean since 1990. All items in the study questionnaire were translated by a team of bilingual (English-Spanish) mental health professionals, including psychiatrists, psychologists, social workers, educators, and epidemiologists. Given known idiomatic differences between countries in Latin America, the translation team was intentionally comprised of representatives of Central, South, and North America, and considerable effort was made to use common language that was understandable to youths in all regions. In cases in which youths in one country were known to be significantly more familiar with a particular idiom or vernacular, minimal exceptions were made for that specific culture to maximize response validity. Back translation was used to ensure that the “content and spirit” of every original item was maintained. Variables The present analysis uses seven dependent variables: lifetime use of alcohol (beer, wine, liquor); lifetime drunkenness; lifetime use of cigarettes; lifetime use of marijuana;


lifetime use of inhalants, tranquilizers, cocaine, crack, ecstasy (which were combined to form an index of number of hard drugs tried); problems with alcohol; and problems with drugs. Lifetime drug use was assessed using items adapted from the Monitoring the Future project [21] and the Drug Use Screening Inventory (DUSI; [7]). Questions specifically asked the respondent “How many times have you consumed the following types of alcohol or drugs in your life?” Specific substances were then listed. The response scale for lifetime use of substances and drinking to drunkenness was 0 (never), 1 (once or twice), 2 (three to four times), and 3 (five or more times). These variables were dichotomized as “never” vs. “other.” Problems with alcohol and Problems with drugs were assessed on 9- and 10-item scales, respectively. Response options ranged from 0 (never) to 3 (five or more times), with higher scores indicating greater problems. Cronbach alpha for alcohol problems was .80 and for drug problems was .87. Six risk factors representing transactions in different microsystems were included in the present study: 1) dysregulation–a deficit in the ability to modulate affect, behavior, and cognition in response to environmental context, resulting in ineffective execution of goal-directed plans (54 items; alpha ⫽ .90), assessed with the dysregulation inventory [6]; 2) family problems with drugs/alcohol (two items); 3) negative family interactions and communication (10 items; alpha ⫽ .74); 4) school disengagement (six items; alpha ⫽ .73); 5) peer deviance (14 items; alpha ⫽ .87); and 6) exposure to violence – witnessing serious violence (five items; alpha ⫽ .82). Family problems with drugs/alcohol, negative family interactions, school disengagement, and peer deviance were assessed with items from the DUSI [7]. Exposure to community violence was assessed with items from the Children’s Report of Exposure to Violence questionnaire (CREV; [11]). Four protective factors representing transactions in different microsystems were included in the present study: 1) belief in God (five items; distinct from spirituality; alpha ⫽ .88); 2) positive family interaction and communication (10 items; alpha ⫽ .82); 3) parent religiosity (four items; alpha ⫽ .76); and 4) positive student-teacher interaction (five items; alpha ⫽ .70). Positive family interaction was assessed with items from the Family Environment Scales [15]. Belief in God and parent religiosity were assessed with items from the Religious Attitudes and Practices Survey [23]. Positive student-teacher interaction was assessed with items from the 2000 Colombian Violence Risk Factor Study [24]. Data Analysis Lifetime substance use was as follows: alcohol use 61.6%; drunkenness 21.8%; cigarette use 40.4%; marijuana use 6.2%; other drug use 18.3%. Problems with alcohol ranged from 0 to 27 (M ⫽ 1.41, SD ⫽ 2.80) and problems with drugs ranged from 0 to 30 (M ⫽ .46, SD ⫽ 2.02).

451

Logistic regressions (for the lifetime substance use outcome variables) and hierarchical linear regression (for the problems with substance use outcome variables) were conducted in two ways. First, analyses were conducted using continuous-level risk and protective factors. Second, risk and protective factors indexes were created based on the distribution of variables in the sample (with values of 1 assigned for participants scoring ⱖ 1 SD above the mean on a variable, and 0 for all other scores), and these indexes were used in the analyses. The risk factor index ranged from 0 to 6 (M ⫽ .96, SD ⫽ 1.24); the protective factor index ranged from 0 to 4 (M ⫽ 1.34, SD ⫽ 1.04). Interactions between risk and protective factors were examined in the model, and follow-up analyses were conducted to identify the specific protective factors that interacted with risk status. Gender, age, and family structure (living with both biological parents vs. not) were controlled in all analyses. In the models predicting lifetime substance use from continuous-level data (Table 1), the risk factors of dysregulation, school disengagement, peer deviance, and exposure to violence most consistently were associated with higher lifetime substance use. Belief in God and parent religiosity were the protective factors most consistently associated with lower lifetime substance use. In the analyses predicting problems with drugs and alcohol from continuous-level data (Table 2), all of the risk factors were negatively associated with problems with alcohol, and most (all but dysregulation and negative family interaction) were associated with problems with drugs. Only belief in God was consistently associated with fewer substance use problems. In the models predicting outcomes from risk and protective factors indexes (Tables 3 and 4), risk factors were associated with increased substance use and substance use problems, and protective factors were associated with decreased substance use and substance use problems across all seven outcomes. Risk and protective factors interacted in predicting lifetime drunkenness and marijuana use, and problems with drugs and alcohol. Tables 5 and 6 present results of analyses disentangling the interactions of risk and protection. For lifetime drunkenness, the risk factor index interacted with belief in God, and with positive family interaction. For lifetime marijuana use, the risk factor index interacted with student-teacher communication (Table 5). For problems with alcohol and problems with drugs, the risk factor index interacted with every protective factor except positive family interaction in the analysis with problems with alcohol as the outcome. The general pattern of the interaction effects is seen in Figure 1. As the figure illustrates, as risk increases, problematic outcomes increase, but protection attenuates that relation to some degree. Luthar et al [19] would call this a reactive effect, where the protective factor has a beneficial effect, but less so at higher levels of risk.

452

W. Kliewer and L. Murrelle / Journal of Adolescent Health 40 (2007) 448 – 455 Table 1 Odds ratios for lifetime substance use by continuous-level risk and protective factors among Central American adolescents OR (p Value)

Control variables Sex Age Family structure Risk factors Dysregulation Fam drug involvement Neg fam interaction School disengagement Peer deviance Exposure to violence Protective factors Belief in God Pos fam interaction Parent religiosity Student-teacher comm.

Alcohol use

Drunkenness

Cigarette use

Marijuana use

Other drug use

1.11*** 1.16*** .87***

1.54*** 1.34*** .83***

1.38*** 1.14*** .92*

1.92*** 1.32*** .90

.65*** 1.07*** 1.00

1.01*** .93** 1.02 1.18*** 1.11*** .99

1.01*** 1.05 1.12 1.15*** 1.11*** 1.03***

1.01*** .99 1.30*** 1.14*** 1.10*** 1.03***

1.00 1.15*** .89 1.10*** 1.15*** 1.06***

1.01*** 1.00 1.40*** 1.04*** 1.04*** 1.06***

1.00 1.04 .92*** .99

.96*** .92 .98* .99

.98** .94 .97*** 1.00

.93*** .89 .97* .97**

.97*** .86*** 1.01 1.01

Fam ⫽ Family; Neg ⫽ Negative; Pos ⫽ Positive; Comm ⫽ Communication. * p ⬍ .05; ** p ⬍ .01; *** p ⬍ .001.

Discussion In the present study, we examined individual risk and protective factors for adolescent substance use and abuse to take advantage of all of the information available in continuous-level variables. We also constructed risk and protective factor indexes to capture the cumulative effects Table 2 Results of regression analyses for problems with substance use by continuous-level risk and protective factors among Central American adolescents Beta (p Value) Problems with alcohol Control variables Sex Age Family structure Risk factors Dysregulation Family drug involvement Negative family interaction School disengagement Peer deviance Exposure to violence Protective factors Belief in God Positive family interaction Parent religiosity Student-teacher communication

Problems with drugs

.05*** .05*** 0

.03*** ⫺.03** ⫺.01

.09*** .05*** .03** .15*** .32*** .06***

.01 .07*** .04 .07*** .24*** .06***

⫺.08*** 0 ⫺.01 0

⫺.13*** ⫺.01 .03*** .01

Note: Beta weights are standardized beta weights from the final step of the regression equation. * p ⬍ .05; ** p ⬍ .01; *** p ⬍ .001.

of risk and protection, and examined interactions between risk and protective factors. Across the seven outcomes indicating substance use and problems with use, several key findings emerged. First, risk factors in multiple domains contributed to substance use and abuse after accounting for age, gender, and family structure. Qualities of the adolescent, the family, the peer culture, and the community environment each added to an adolescent’s risk for substance use and associated difficulties. In the individual domain, dysregulation, which refers to a deficit in the ability to modulate affect, behavior, and cognition in response to environmental context [6], was a strong risk factor for substance use and problems with alcohol. Dysregulation may interact with family and peer factors in adolescence by evoking particular responses from the environment, thereby leading to substance use [25]. The findings with dysregulation in the current study mirror results with adolescents in the United States [6]. Adolescents’ disengagement from school also independently contributed to substance use and problems with use. This finding is consistent with a large body of literature (c.f. [10,17]) demonstrating that lack of academic interest or motivation pulls adolescents toward deviant peers, and heightens risk for externalizing behavior problems, including substance use. In the peer domain, being friends with deviant peers contributed to substance use and problems with use beyond contributions in other domains. This finding is consistent with many studies highlighting the contributions of peers to externalizing behavior problems [9]. Peers provide access to drugs and alcohol, and contribute to difficulties adolescents experience with substance use by contributing to poor decision-making.


453

Table 3 Odds ratios for lifetime substance use by risk and protective factors indexes and their interactions among Central American adolescents OR (p Value)

Control variables Sex Age Family structure Risk factor index Protective factor index Interaction of risk and protective factors

Alcohol use

Drunkenness

Cigarette use

Marijuana use

Other drug use

1.12*** 1.24*** .90** 1.47*** .69*** 1.01

1.64*** 1.38*** .88** 1.76*** .74*** 1.05**

1.43*** 1.20*** .94 1.57*** .76*** 1.01

2.28*** 1.35*** .95 1.93*** .65*** 1.05*

.67*** 1.09*** .98 1.50*** .89** 1.01

* p ⬍ .05; ** p ⬍ .01; *** p ⬍ .001.

The findings associated with exposure to violence add to a developing literature from investigators around the world linking victimization or witnessing violence to drug use [8,12,16]. Stress relief is one of the main reasons adolescents who are exposed to violence use drugs and alcohol. Notably, exposure to violence in the present study contributed to substance use beyond the contributions of demographic factors, and risk factors in the individual, family, and peer domains. Negative family interactions predicted cigarette and other drug use as well as problems with alcohol, but not alcohol or marijuana use, or problems with drugs. Shared variation between family structure or family drug involvement and the measure of negative family interaction may have accounted for these findings. In terms of protective factors, a personal belief in God (in contrast to mere religious attendance, or more general spirituality) and parent religiosity were most consistently related to lower lifetime substance use and substance use problems. Adolescents who espouse a belief in a God who is personal and who desires a relationship with them were much less likely than their peers who did not espouse such a belief to use substances or report problems with use. Perhaps the attachment to God and resultant desire to live in a way that pleases God accounts for these associations. Table 4 Results of regression analyses for problems with substance use by risk and protect factor indexes and their interactions among Central American adolescents Beta (p Value)

Control variables Sex Age Family structure Risk factor index Protective factor index Interaction of risk and protective factors * p ⬍ .05; ** p ⬍ .01; *** p ⬍ .001.

Problems with alcohol

Problems with drugs

.07*** .10*** .01 .38*** ⫺.12*** ⫺.09***

.06*** .01 0 .27*** ⫺.07*** ⫺.13***

These data are consistent with research showing that adolescents with strong attachments to God engage in low levels of risk-taking behavior [13]. Parents who are religious may also model healthier lifestyles or be more actively engaged in authoritative parenting [14], which is protective for adolescents and could account for these findings. One intention of the present study was to examine the interactions of risk and protective factors, and to disentangle any significant interaction effects. There were significant risk-factor-by-protective-factor index interactions for four of the seven study outcomes: lifetime drunkenness, lifetime marijuana use, problems with drugs, and problems with alcohol. In disentangling these interactions, different protective factors emerged as salient for different outcomes related to use, but all protective factors were important for problems associated with drug and alcohol use. Table 5 Odds ratios for lifetime drunkenness and lifetime use of marijuana: disentangling risk and protective factors interactions among Central American adolescents OR (p Value)

Control variables Sex Age Family structure Risk factor index Protective factors Belief in God Pos fam interaction Parent religiosity Student-teacher comm. Interactions Risk ⫻ belief in God Risk ⫻ pos fam interaction Risk ⫻ parent religiosity Risk ⫻ student-teacher communication

Drunkenness

Marijuana use

1.63*** 1.38*** .89** 1.77***

2.19*** 1.36*** .94 1.91***

.94*** .89** .95*** .96***

.92*** 1.02 .94*** .93***

1.01* 1.13*** 1.00 1.00

1.00 1.05 1.00 1.02*

Fam ⫽ Family; Neg ⫽ Negative; Pos ⫽ Positive; Comm ⫽ Communication. * p ⬍ .05; ** p ⬍ .01; *** p ⬍ .001.

454


Table 6 Results of regression analyses for problems with alcohol and drugs: disentangling risk and protective factors interactions among Central American adolescents Beta (p Value) Problems with alcohol Control variables Sex Age Family structure Risk factor index Protective factors Belief in God Pos fam interaction Parent religiosity Student-teacher comm. Interactions Risk ⫻ belief in God Risk ⫻ pos fam interaction Risk ⫻ parent religiosity Risk ⫻ student-teacher communication

Problems with drugs

.07*** .10*** .01 .38***

.06*** .01 0 .24***

⫺.07*** 0 ⫺.05*** ⫺.04***

⫺.07*** 0 0 ⫺.02**

⫺.07*** .01 ⫺.05*** ⫺.02**

⫺.16*** ⫺.06*** ⫺.03** ⫺.02***

* p ⬍ .05; ** p ⬍ .01; *** p ⬍ .001.

For lifetime drunkenness, a personal belief in God and positive family interaction attenuated risk. Consistent with prior research documenting the protective effects of a personal belief in God [13] and parental monitoring and support [16,17], as risk levels increased, adolescents with high levels of positive family interaction or a personal belief in God were less likely to report ever having been drunk. For lifetime marijuana use, positive student-teacher interaction

attenuated risk. This finding reflects the importance of extrafamilial support during adolescence [17,18] and suggests that one strategy to lower risky substance use in adolescence is to foster strong, positive bonds between adolescents and their teachers. Limitations Although the sample for this study was large, the sample was restricted to adolescents attending school, and to adolescents residing in Costa Rica, Guatemala, and Panama. Therefore, the findings cannot be generalized to all adolescents living in Central America. The data on which the study was based were correlational and collected at one point in time, thus, causation should not be inferred from these findings. Conclusion These data suggest that substance use among Central American adolescents is of concern, and highlight the importance of assessing risk and protective factors across multiple ecologies of adolescents’ lives. Further, prevention and intervention efforts need to target risk and protective factors in each of the domains in the present study: individual, family, peer, school, and community. For example, these data suggest that the governments of Costa Rica, Guatemala, and Panama should continue their policies aimed at reducing community violence. At the family level, interventions might be implemented to both reduce negative interaction patterns and increase positive interactions

Figure 1. General pattern of interactions between the risk and protective factors indicies, illustrated here with the dependent variable of problems with alcohol.


among family members. This might involve training family members in conflict resolution and communication skills, and encouraging families to establish regular times for relaxed, positive activities. This type of family intervention might have the added benefit of reducing involvement with deviant peers. In the school domain, interventions might be implemented during elementary school to identify and treat children with learning difficulties, thereby reducing their likelihood of school disengagement. Once children reach secondary school, additional interventions might be implemented both with adolescents and with teachers to enhance opportunities for connections with teachers. Finally, at the individual level, interventions are needed that address deficits in emotion regulation skills. We believe that this type of large-scale, cross-cultural, population-based study of adolescents underscores the feasibility of basing national and regional policies aimed at promoting adolescent health and preventing adolescent problem behaviors on data collected within the same cultural context. Prevention experts and public health policy makers are best served by having at hand such “local” information as critical allocations are made in the context of always-limited resources.

Acknowledgment We thank the members of the Inter-American Workgroup on Dysregulation and Prevention of Drug Abuse for their contributions to this research project.

References [1] Chen C, Dormitzer CM, Gutierrez U, et al. The adolescent behavioral repertoire as a context for drug exposure: behavioral autarcesis at play. Addiction 2004;99:897–906. [2] Bronfenbrenner U. Contexts of child rearing: problems and prospects. Am Psychol 1979;34:844 –50. [3] Blum RW, Ireland M. Reducing risk, increasing protective factors: findings from the Caribbean Youth Health Survey. J Adolesc Health 2004;35:493–500. [4] Masten A. Ordinary magic. Resilience processes in development. Am Psychol 2001;56:227–38. [5] Rutter M. Psychosocial resilience and protective mechanisms. Am J Orthopsychiatry 1987;57:316 –21. [6] Mezzich AC, Tarter RE, Giancola PR, Kirisci L. The dysregulation inventory: a new scale to assess the risk for substance use disorder. J Child Adolesc Subst Abuse 2001;10:35– 43. [7] Tarter RE, Kirisci L, Mezzich A. The Drug Use Screening Inventory: school adjustment correlates of substance abuse. Meas Eval Couns Dev 1996;29:24 –34.

455

[8] Kilpatrick DG, Acierno R, Saunders B, et al. Risk factors for adolescent substance abuse and dependence: data from a national sample. J Consult Clin Psychol 2000;68:19 –30. [9] Patterson GR, DeBaryshe BD, Ramsey E. A developmental perspective on antisocial behavior Am Psychol 1989;44:329 –35. [10] Kasen S, Cohen P, Brook JS. Adolescent school experiences and dropout, adolescent pregnancy, and young adult deviant behavior. J Adolesc Res 1998;13:49 –72. [11] Cooley-Quille M, Turner SM, Biedel DC. Emotional impact of children’s exposure to violence: a preliminary study. J Am Acad Child Adolesc Psychiatry 1995;34:1362– 8. [12] Vermeiren R, Schwab-Stone M, Deboutte D, et al. Violence exposure and substance use in adolescents: findings from three countries. Pediatrics 2003;111:535– 40. [13] Kliewer W, Wade NG, Worthington EL Jr. Religion and spirituality in adolescents: preventing dysfunction and promoting health. In: Gullotta TP, Bloom M, eds. The Encyclopedia of Primary Prevention and Health Promotion. New York: Plenum, 2003:881– 8. [14] Gunnoe ML, Hetherington EM, Reiss D. Parental religiosity, parenting style, and adolescent social responsibility. J Early Adolesc 1999; 19:199 –225. [15] Moos R, Moos B. Family Environment Scale Manual, 2nd ed. Palo Alto, CA: Consulting Psychologists Press, 1994. [16] Sullivan TN, Kung EM, Farrell AD. Relation between witnessing violence and drug use initiation among rural adolescents: parental monitoring and family support as protective factors. J Clin Child Adolesc Psychol 2004;33:488 –98. [17] Resnick MD, Bearman PS, Blum RW, et al. Protecting adolescents from harm: findings from the National Longitudinal Study on Adolescent Health. JAMA 1997;278:823–32. [18] Cattley G. The impact of teacher-parent-peer support on students’ well-being and adjustment to the middle years of schooling. Int J Adolesc Youth 2004;11:269 – 82. [19] Luthar SS, Cicchetti D, Becker B. The construct of resilience: a critical evaluation and guidelines for future work. Child Dev 2000; 71:543– 62. [20] Murrelle L. Progress Report on the Study on Psychological Dysregulation as a Risk Factor for Drug Abuse, Violent Behaviors, and Social Adjustment among Central American Youths. Richmond, VA: Virginia Commonwealth University, 2001. [21] Bachman JG, Johnston LD, O’Malley PM. Monitoring the Future Project After Twenty-Two Years: Design and Procedures (Monitoring the Future Occasional Paper 38). Ann Arbor, MI: The University of Michigan, Institute for Social Research, 1996. [22] Murrelle L, Prom E, Aggen S, et al. Psychological dysregulation and drug involvement among Panamanian youths. Poster presented at the annual meeting of the College on Problems of Drug Dependence, Scottsdale, AZ, 2001. [23] D’Onofrio BM, Eaves LJ, Murrell L, et al. Understanding biological and social influences on religious affiliation, attitudes, and behaviors: a behavior genetic perspective. J Pers 1999;6:953– 84. [24] Maya JM, Torres Y, Murrelle L, et al. Use and Abuse of Psychoactive Substances and Associated Factors in Juvenile Delinquency. Medellin, Colombia: Institute for Health Sciences, Colciencias and InterAmerican Development Bank, 2000. [25] Dawes MA, Antelman, SM, Vanyukov MM, et al. Developmental sources of variation in liability to adolescent substance use disorders. Drug Alcohol Depend 2000;61:3–14.