Effects of the Universal Prevention Curriculum for Schools on Substance Use Among Peruvian Adolescents: A Randomized Trial

Abstract

This group-randomized trial assessed the effects of a universal prevention training curriculum for school administrators and teachers that focused on effective strategies to prevent adolescent substance use and related problems. Twenty-eight schools in three regions of Peru were randomly assigned to either an intervention or control condition (14 schools per condition). Repeated cross-sectional samples of 11 to 19-year-old students participated in four surveys from May 2018 to November 2019 (N=24,529). School administrators and teachers at intervention schools participated in a universal prevention training curriculum focusing on the development of a positive school climate as well as effective policies related to school substance use. All intervention and control schools were offered Unplugged, a classroom-based substance use prevention curriculum. Outcome measures included: lifetime drug use; past-year and past-month tobacco, alcohol, marijuana, and other drug use; awareness of school tobacco and alcohol use policies; perceived enforcement of school policies; school bonding; perceived friends’ use of tobacco, alcohol, marijuana and other drugs; and personal problems in general and problems related to substance use. Multi-level analyses indicated significant reductions in past-year and past-month smoking, friends’ substance use, and problems related to substance use and in general at intervention relative to control schools. Significant increases were found in intervention vs. control schools related to students’ awareness of school substance use policies, perceived likelihood of getting caught for smoking, and school bonding. These findings suggest that the universal prevention training curriculum and the school policy and climate changes it promoted reduced substance use and related problems in the study population of Peruvian adolescents.

Introduction

Data from the World Health Organization indicate that before the COVID-19 pandemic began in 2020, non-communicable diseases (NCDs), primarily ‘man-made’ or behavioral health problems related to pollution, diet, and substance use, constituted 71%, or 40.5 million of the 56.9 million global deaths (World Health Organization, 2018). Among the 34 risk factors associated with deaths worldwide, smoking is ranked second, alcohol use ninth, secondhand smoke fifteenth, and other drug use twenty-fourth. These substance use risk factors are of particular concern in low- and middle-income countries, where 85% of premature deaths occur from NCDs (WHO, 2018).

To address this concern, Applied Prevention Science International staff developed the Universal Prevention Curriculum (UPC) series in collaboration with U.S. prevention researchers who are experts in substance use epidemiology and prevention strategies that are delivered to families; within schools, the workplace, and the community; through the media; and by means of regulatory policies. The UPC consists of two parallel training programs designed to support the development of substance use prevention professionals (International Society of Substance Use Professionals, 2021). The first series is designed for prevention coordinators and is intended to increase the knowledge base of decision makers, supervisors, and planners of substance use prevention programming. The second series, which targets implementers, was designed to improve the competencies and skills of front-line professionals who administer prevention interventions or policies. It was expected that these implementers would be supervised by the prevention coordinators, whose training should precede that of their front-line staff. The UPC was also designed to highlight salient findings from prevention science and its application to the selection of evidence-based prevention interventions and policies, and to focus attention on the infrastructure required at the community or organization level to support and sustain these strategies over time. The UPC is grounded in the International Standards on Drug Use Prevention (United Nations, 2018) and draws on the European Drug Prevention Quality Standards, which specify the competencies and skills required by prevention professionals to deliver interventions and policies effectively (European Monitoring Centre on Drugs and Drug Addiction, 2011).

Both UPC series are structured around a number of critical themes such as the definition of substance use, the science of prevention, evidence-based prevention interventions and policies, the developmental nature and etiology of substance use, critical theories underlying evidence-base prevention interventions, professional competencies and skills, and prevention ethics. Both series present three interdependent components relating to evidence-based prevention interventions and polices. These comprise content (key knowledge and information such as social, decision-making, and communication skills), structure (the number of sessions involved), and delivery (appropriate implementation and instructional strategies). The series also emphasize strategies that have been found not to be effective, such as didactic lecturing of adolescents rather than using interactive teaching techniques.

In collaboration with the University of Peru Cayetano Heredia (UPCH) in Lima, we conducted an evaluation of the effects of UPC’s School-based Prevention training curriculum which is designed for substance use prevention in school settings and is based on the Protective Schools model (Bosworth, 2000). The curriculum presents an overview of research findings from the prevention science field that supports school-based prevention interventions and policies, as well as strategies designed to improve school climate. In addition, the curriculum also provides guidance as to how to monitor and sustain these various strategies over time. Further information concerning the content of the training curriculum may be obtained upon request from the authors.

We conducted our study in Peru, a middle-income country (World Bank Country and Lending Groups, 2021) where the 2017 National Survey on Secondary Level Students indicates that both legal and illegal drug use by adolescents remain prevalent (DEVIDA, 2019). For example, annual prevalence rates for either tobacco or alcohol use were 9.8% among 11 to 13-year-olds, 27.3% among 14 to 16-year-olds, and 35.4% among 17 to 20-year-olds, while annual prevalence rates for any illegal drug use (marijuana, cocaine, PBC, inhalants or Ecstasy) in these three age groups were 3.3%, 5.3%, and 9.3%, respectively. Epidemiologic and survey data consistently indicate that the use of psychoactive substances is driven by normative beliefs and favorable attitudes toward these substances. The 2017 national survey results show that about 25% of Peruvian secondary school students had positive attitudes toward substance use.

We hypothesized that secondary schools participating in the UPC training would demonstrate improvements in the adoption and implementation of school policies prohibiting alcohol and tobacco use, and creating a positive school climate for students. We also hypothesized that the likelihood of substance use behaviors, related problems, and perceived substance use among friends would decrease among students at intervention relative to control schools, while awareness of school alcohol and tobacco policies, perceived policy enforcement, and school bonding would increase among students at intervention relative to control schools.

Methods

Study Design

This study used a group randomized controlled design to evaluate the UPC training and related school intervention with 28 secondary schools in three Peru regions (Lima, Callao, Cajamarca). The number of schools to be included was determined by a power analysis that assumed six repeated biannual cross-sectional surveys across a projected three-year period (2018–2020). However, due to the COVID-19 pandemic and consequent school closings in Peru, study data were limited to four waves collected from 2018 and 2019.

School eligibility criteria included: secondary school grades 1–5 (ages 11 to 20), no substance use prevention or related program, and willingness to participate in a controlled study with random assignment to intervention or control group. These criteria were shared with specialists from the Regional Directorates of Education (DREs) of Lima, Callao and Cajamarca, who provided the UPCH project team with a list of candidate schools in each city. These DREs then invited approximately 30 schools in each region to participate in a meeting to determine their suitability for and willingness to participate in the study. The UPCH team visited each school and interviewed its principal to make a final determination of the suitability of the school for the study. Teachers in all eligible schools were given the opportunity to receive training in Unplugged (Kreeft et al., 2009), a classroom-based substance use prevention curriculum that was developed in Europe for children 10–14 years of age. This theory-based curriculum targets a number of proximal outcomes that are predictive of substance use, including risk perceptions, attitudes toward drugs, normative beliefs, critical and creative thinking, refusal skills, and problem solving and decision making. It has been evaluated in several countries and has yielded positive outcomes, particularly related to alcohol use (Agabio et al., 2015; Vigna-Taliente et al., 2014).

All schools administered a substance use survey to their students, and results were used by the evaluation team to select, at random, 28 schools within the three regions from the eligible pool for the study, yielding 14 in each condition. Selected schools were matched into pairs within each region based on similarity of: (a) percent male; (b) socioeconomic status of the school catchment area; and (c) prevalence of past year substance use behaviors. Matched schools were randomly assigned using a computerized algorithm to either the UPC intervention or control condition. Figure 1 displays a CONSORT diagram that displays the flow of schools and students.

Fig 1.

CONSORT diagram

UPC Training and Intervention Implementation

UPC training of school administrators and teachers occurred in April 2018. The principals of schools in the intervention group, and the teachers these principals selected, participated in an introductory UPC course comprising topics related to the foundations of prevention, basic practitioner skills, and an overview of the science of evidence-based prevention interventions and policies. This course was followed by a second curriculum that addressed issues and concepts specifically related to substance use prevention in school contexts. This curriculum included guidance to school staff as to establishing prevention leadership and action teams designed to develop an organizational structure for prevention activities. It was followed by instruction concerning the selection and implementation of effective school prevention policies and the promotion of a healthy, supportive, and engaging school climate.

After the training, PLATs at the 14 intervention schools began to develop alcohol and tobacco use policy and school climate improvement activities. Policy awareness and school climate improvement activities were initiated in the fall of 2018, and Unplugged curriculum training was provided to all teachers. All school intervention activities were fully implemented from March through December 2019.

Observers from the UPCH project team rated the uptake and implementation of intervention activities in each participating intervention school that corresponded with the key UPC training components: (1) school policy, (2) school climate, (3) the Unplugged prevention curriculum, and (4) PLAT activities. During the 2018 and 2019 school years, three raters monitored the required elements for each component at least twice monthly at each intervention school using a four-point scale (1 = did not achieve objectives; 4 = Completed all requirements and component-related actions in an optimal manner). Note that Unplugged and related staff training were offered to all control schools, and the degree to which the curriculum was implemented in these schools was monitored.

Student Surveys

Self-administered, anonymous, cross-sectional student surveys were administered by members of the UPCH research team in all classrooms for grades 1 through 5 at participating schools in May 2018, December 2018, June 2019, and November 2019. One school in the control condition declined to permit survey administration in December of 2018 for reasons unrelated to this study. Average survey response rates for the four data collection periods ranged from 87.4% to 90.6%. Of the 27,012 students who completed surveys across the four surveys waves, 24,529 (91%) provided complete data for all study variables.

Survey Measures

All survey questions were in Spanish. Some questions were only included in the first three surveys. Unless otherwise noted, questions were included in all four surveys.

Substance use.

Students were asked if they had ever (a) smoked cigarettes; (b) drunk alcoholic beverages; or used (c) marijuana, (d) inhalants, (e) cocaine paste, (f) cocaine, (g) amphetamines or methamphetamines, (h) hallucinogens (e.g., PCP, LSD), and (i) San Pedro, Ayahuasca, huachuma, or peyote. Students who responded affirmatively to the questions for smoking cigarettes, drinking alcoholic beverages, and marijuana and inhalant use were asked if they had used these substances in the past 12 months and, if so, whether they had used these substances in the past 30 days. Students were also asked about the frequency of using these substances within the past 12 months and 30 days. Because of the skewed response distributions and low prevalence rates for marijuana and inhalant use (< 4%), dichotomous measures for smoking cigarettes and drinking alcoholic beverages in the past 12 months and past 30 days were created. Because prevalence rates for lifetime use of marijuana, inhalants, and other illegal drugs were also very low (< 6%), we created an overall measure of any lifetime use of marijuana, inhalants, or other drugs. Students were also asked how often they consumed five or more alcoholic drinks on one occasion in the past two weeks, with five response options that ranged from “none” to “more than five times”. Because of the skewed response distribution, we created a dichotomous variable indicating any binge drinking in the past two weeks.

Perceived substance use among friends.

In the first three surveys, students were asked how many friends who they spent most of their time with (a) smoke cigarettes, (b) get drunk, and (c) use marijuana or other drugs. Four possible responses were “none,” “any of them,” “most of them,” and “all of them”. These were coded on a four-point scale with higher values indicating a greater proportion of friends who engage in each of these substance use behaviors. We also created a summative measure with these three variables, in which higher values represented a greater proportion of friends who engaged in any substance use (α = 0.83).

School substance use policies and perceived enforcement.

Students were asked whether they were aware of rules that prohibit (a) smoking and (b) alcohol use on or near the school (yes/no). They were also asked how likely it is that a student would get into trouble for (a) smoking and (b) alcohol use at school with four possible responses ranging from “very unlikely” to “very likely” that were coded 1–4.

School bonding.

Students were asked how much they agreed or disagreed with the following statements: “Students in my class like to be together;” “Most people in my class are kind and help you;” “The other students accept me as I am;” and “I have a lot of respect for what teachers tell me.” Four possible responses to each statement were “strongly agree,” “agree,” “disagree,” “strongly disagree”. Response values were reverse coded, and the mean response was then computed to represent school bonding (α = 0.73).

Problems related to substance use.

Students were asked whether they experienced 11 different problems related to alcohol or other drug use in the past 12 months, including (a) scuffle or argument, (b) fight, (c) accident or injury, (d) loss of money or valuables, (e) damage to clothing or other personal belongings, (f) relationship problems with parents, (g) relationship problems with friends, (h) relationship problems with teachers, (i) poor school performance, (j) victim of theft or robbery, and (k) hospitalized or rushed to emergency room. Responses to these questions were summed to create an index of problems related to substance use. In the first three surveys, students were also asked if they experienced each of these problems for any other reason. We used data from the first three surveys to create an index of problems related to substance use or other reasons.

Demographic characteristics.

Students reported their age and sex (male, female).

Data Analysis

We conducted t-tests to compare intervention implementation ratings in the 14 intervention schools in 2018 and 2019. Based on the timing of intervention activities and statistical power considerations, we combined data from the May and December 2018 cross-sectional surveys (pre-full intervention implementation) and the June and November 2019 cross-sectional surveys (full intervention implementation period) for analyses in 2020. We first examined descriptive statistics for all variables in the 2018 and 2019 survey years for the total sample and by intervention condition. Chi-square and t-tests were conducted to assess differences between 2018 and 2019 for these variables. Multi-level regression analyses were conducted to assess UPC training and intervention effects on substance use and related outcomes targeted by the intervention. Logistic and linear regression analyses were conducted for binary and continuous outcomes, respectively. Regression models included the following at each level: student (level 1) outcome variables, a dummy variable representing pre- and post-intervention years, and demographic characteristics (age, gender); a school (level 2) intervention condition dummy variable, and a cross-level interaction term for intervention condition × pre-post intervention years, and random effects for outcome intercepts and slopes; and a city (level 3) random effect for outcomes. Random effects accounted for variance in outcomes attributable to students nested within schools, which were nested within cities, and the slope random effects accounted for unexplained variation in outcome slopes over time. HLM version 8.0 software was used for all analyses (Raudenbush et al., 2019). This modeling approach has been used in a number of previous group randomized intervention trials involving repeated cross-sectional data from middle, high school, and college students (e.g., Flewelling et al., 2013; Paschall et al., 2011a,b; Saltz et al., 2010).

Results

Intervention Implementation

As shown in Table 1, there was a statistically significant improvement in overall intervention implementation from 2018 to 2019. Analyses focusing on the four domains indicated overall improvement with school policy and school climate, while the implementation scores for 2018 and 2019 were not significantly different for the Unplugged curriculum. We also obtained information from intervention and control schools regarding students’ exposure to the Unplugged curriculum (i.e., numbers of classes and students that participated in 2018 and 2019) and found similar levels of exposure between intervention and control schools. Thus, any observed differences in outcomes among students at intervention relative to control schools would likely be attributable to improvements in school policy and climate.

Table 1.

Comparison of fidelity domain ratings for the 14 intervention schools in 2018 and 2019, mean (standard deviation)

Domain	2018	2019
School policy	3.00 (.57)	3.54 (.64)*
School climate	2.95 (.39)	3.52 (.36)**
School curriculum	2.95 (.71)	3.22 (.70)
Prevention leadership action team	2.96 (.66)	3.39 (.67)
Overall rating	2.96 (.52)	3.42 (.54)*

^*p < .05,

^**p < .01

Sample Characteristics

Descriptive statistics in Table 2 indicate that the mean student age was approximately 14 and about half of the students were male in 2018 and 2019. Small though significant changes in these demographic characteristics were observed from 2018 to 2019.

Table 2.

Sample characteristics by intervention condition and year, mean (SD) or percent

Variable	Total sample		Intervention schools		Control schools
	2018 (N=11,148)	2019 (N=13,381)	2018 (n=5,681)	2019 (n=6,684)	2018 (n=5,467)	2019 (n=6,697)
Demographic characteristics
Age	14.2 (1.6)	14.3 (1.5)**	14.2 (1.6)	14.3 (1.5)*	14.2 (1.6)	14.3 (1.5)*
Male (%)	50.5	51.8**	49.3	51.0**	51.8	52.6
Substance use behaviors
Past-year smoking (%)	14.0	10.4**	13.3	8.7**	14.7	12.1**
Past-month smoking (%)	9.7	6.5**	9.4	5.3**	10.0	7.8**
Past-year alcohol use (%)	22.2	20.3**	20.6	17.5**	23.8	23.0**
Past-month alcohol use (%)	12.7	11.0**	11.7	9.2**	13.8	12.9**
Binge drinking, past 2 weeks (%)	7.7	7.5	7.3	6.2*	8.1	8.7
Lifetime drug use (%)a	10.4	7.9**	10.5	7.7**	10.4	8.1**
Past-year substance use (%)b	26.0	23.0**	24.4	20.0**	27.7	25.9**
Past-month substance use (%)b	16.1	13.6**	14.9	11.5**	17.3	15.8**
Friends’ substance use c
Smoking	1.5 (0.7)	1.4 (0.7)**	1.5 (0.7)	1.4 (0.7)**	1.5 (0.7)	1.5 (0.7)
Getting drunk	1.6 (0.8)	1.6 (0.8)	1.6 (0.8)	1.5 (0.7)**	1.6 (0.8)	1.6 (0.8)
Marijuana or other drugs	1.3 (0.6)	1.3 (0.6)	1.3 (0.6)	1.2 (0.6)*	1.3 (0.6)	1.3 (0.6)
Total friends’ substance use	4.3 (1.8)	4.2 (1.8)**	4.3 (1.8)	4.2 (1.7)**	4.3 (1.8)	4.4 (1.8)
School policies, bonding
Smoking is prohibited (%)	69.7	66.3**	70.5	68.9	69.0	63.7**
Alcohol use is prohibited (%)	57.7	58.3	58.4	61.5**	56.9	55.1
Likelihood of getting into trouble for smoking at school	2.5 (1.3)	2.7 (1.3)**	2.5 (1.3)	2.6 (1.3)**	2.6 (1.3)	2.7 (1.3)**
Likelihood of getting into trouble for alcohol use at school	2.6 (1.2)	2.7 (1.2)*	2.6 (1.2)	2.6 (1.2)	2.6 (1.2)	2.7 (1.2)
School bonding	3.2 (0.5)	3.1 (0.6)	3.1 (0.6)	3.2 (0.6)	3.2 (0.5)	3.1 (0.6)**
Problems in the past year
Problems related to substance use	0.4 (1.7)	0.7 (2.0)**	0.4 (1.6)	0.7 (1.8)**	0.4 (1.7)	0.8 (2.0)**
Problems for any reasonc	2.5 (3.2)	2.5 (3.3)	2.3 (3.1)	2.2 (3.1)**	2.6 (3.3)	2.8 (3.4)

^*p < .05,

^**p < .01

^aMarijuana, inhalants, or other drugs (e.g., cocaine, hallucinogens).

^bTobacco, alcohol, marijuana, or inhalants.

^cBased on both 2018 surveys and the June 2019 survey (N=6,806; intervention schools n=3,401; control schools n=3,405).

Substance use.

Significant reductions in past-year and past-month smoking and alcohol use were observed for the total sample and for both intervention and control schools from 2018 to 2019. For example, among students at intervention schools, past-year smoking prevalence decreased from 13.3% to 8.7%, and from 14.7% to 12.1% among students in control schools. There were also significant reductions in lifetime drug use and overall measures of past-year and past-month substance use for the total sample and among students in both intervention and control schools.

Perceived substance use among friends.

Significant reductions in the proportion of friends who were perceived to smoke, get drunk, and use marijuana or other drugs were observed for the total sample and among intervention students 2018 to 2019. No significant changes were seen in the control schools.

School policies and perceived enforcement.

From 2018 to 2019, there was a significant decrease in the overall percentage of students who were aware of their school’s no-smoking policy. However, among intervention school students, there was no significant change in awareness of the school’s tobacco use policy (~70%), while there was a significant decrease among control school students (69.0% to 63.7%). There was a significant increase from 2018 to 2019 in the percentage of intervention students who were aware of their school’s no-alcohol use policy (58.4% to 61.5%), but no significant changes for the total sample or among control students. There were significant increases in the perceived likelihood of getting into trouble at school for smoking on or near campus for the total sample and among intervention and control students. There was also a significant increase from 2018 to 2019 in the perceived likelihood of getting into trouble at school for alcohol use on or near campus for the total sample, but no significant changes among intervention and control students.

School bonding.

There were no significant changes in the level of school bonding from 2018 to 2019 for the total sample or among students at intervention schools. However, there was a significant decrease in school bonding among students in control schools.

Problems related to substance use.

There were significant increases from 2018 to 2019 in the number of problems related to substance use reported by students in the total sample or those at intervention and control schools. In contrast, there was a significant decrease from 2018 to 2019 in the number of problems experienced for any reason among students at intervention schools, but no significant changes for the total sample or among students in control schools.

Multi-Level Regression Analyses

Results of multi-level logistic regression analyses are in Table 3. For substance use behaviors, UPC training and associated school-related interventions significantly affected past-year and past-month smoking. Among students in intervention versus control schools, there was a 20% relative reduction in the odds of past-year smoking from 2018 to 2019 and a 33% relative reduction in the odds of past-month smoking from 2018 to 2019. The relative change in past-month smoking is shown in Figure 2. The analyses indicated that there were relative reductions in the odds of those behaviors from 2018 to 2019 among students in intervention versus control schools, ranging from an 8% reduction for lifetime drug use to a 22% reduction for binge drinking in the past two weeks. However, although substantively large, these effects were not statistically significant.

Table 3.

Summary of results from multi-level logistic regression analyses to assess UPC training and intervention effects on substance use behaviors and school policy awareness

Outcome	Odds Ratio (95% confidence interval)a
Substance use behaviors
Past-year smoking	0.80 (0.65, 0.99)*
Past-month smoking	0.67 (0.51, 0.92)*
Past-year alcohol use	0.88 (0.74, 1.06)
Past-month alcohol use	0.84 (0.65, 1.09)
Binge drinking, past 2 weeks	0.78 (0.53, 1.15)
Lifetime drug useb	0.92 (0.67, 1.27)
Past-year substance usec	0.88 (0.74, 1.03)
Past-month substance usec	0.83 (0.66, 1.04)
School policies
Smoking is prohibited	1.23 (1.02, 1.48)*
Alcohol use is prohibited	1.29 (1.08, 1.53)**

^*p < .05,

^**p < .01

^aOdds ratio for pre/post intervention year × intervention condition, adjusted for student age and gender.

^bMarijuana, inhalants, or other drugs (e.g., cocaine, hallucinogens).

^cTobacco, alcohol, marijuana, or inhalants.

Fig 2.

Change in the prevalence of past-month smoking among students by intervention condition.

Regarding awareness of school substance use policies, there were significant increases in students’ awareness of school tobacco and alcohol use policies among students in intervention versus control schools. From 2018 to 2019, there was a 23% increase in the odds that students at intervention schools, relative to those at control schools, would be aware of their school’s tobacco use policy. Similarly, there was a 29% increase from 2018 to 2019 in the odds that students at intervention schools, relative to those at control schools, would be aware of their school alcohol use policy. The relative change in students’ awareness of their school’s alcohol use policy is shown in Figure 3.

Fig 3.

Change in students’ awareness of their schools’ policy prohibiting alcohol use by intervention condition.

Results of multi-level linear regression analyses in Table 4 indicate a non-significant, although suggestive, increase (p < .10) from 2018 to 2019 in the perceived likelihood of getting into trouble at school for smoking among students at intervention relative to control schools. There were no significant differences in the perceived likelihood of getting into trouble at school for drinking alcoholic beverages. There was also a suggestive, but non-significant (p < .10) increase in the level of school bonding from 2018 to 2019 among students at intervention versus control schools.

Table 4.

Summary of results from multi-level linear regression analyses to assess UPC training and intervention effects on psychosocial outcomes and problems

Outcome	beta (standard error)a
Perceived school policy enforcement
Likelihood of getting into trouble for smoking at school	0.06 (.03)†
Likelihood of getting into trouble for alcohol use at school	0.03 (.04)
School bonding	0.05 (.03)†
Friends’ substance use
Smoking	−0.04 (.03)
Getting drunk	−0.05 (.03)
Marijuana or other drugs	−0.03 (.02)
Total friends’ substance use	−0.13 (.07)†
Problems in the past year
Problems related to substance use	−0.19 (.07)*
Problems for any reason	−0.25 (.12)*

^*p < .05,

^†p < .10

^aBeta coefficient for pre/post intervention year × intervention condition, adjusted for student age and gender.

There were no significant differences in the perceived proportion of friends who smoked cigarettes, got drunk, or used marijuana or other drugs from 2018 to 2019 among students in intervention versus control schools. However, there was a reduction in the total perceived proportion of friends who used any of these substances from 2018 to 2019 among students in intervention versus control schools, although it did not reach a traditional level statistical significance (p < .10).

Significant reductions in the number of problems related to substance use and problems experienced for any reason in the past year were observed among students from 2018 to 2019 in intervention schools, relative to control schools. The relative change in problems experienced in the past year for any reason is shown in Figure 4.

Fig 4.

Change in problems experienced by students in the past year by intervention condition.

Discussion

This group randomized trial conducted in Peru constitutes the first formal outcome evaluation of any of the Universal Prevention Curriculum (UPC) series. The focus of our evaluation, the curriculum designed for schools, included school leadership team development, the development and dissemination of appropriate school policies regarding substance use, the promotion of a positive school climate, and the delivery of an evidence-based school prevention curriculum. The findings suggest that the UPC training was associated with an increased likelihood that evidence-based prevention programming, particularly related to school policy and climate, would be implemented. The improvements in school policy and climate contributed to significant reductions in several key outcomes related to adolescent substance use. The most salient of these was students’ tobacco use, a finding that is particularly important not only because of the well-documented link between tobacco use and mortality and morbidity, but also because adolescent tobacco use is associated with later use of marijuana and other drugs (Lindsay & Rainey, 1997; Torabi et al., 2010; Tullis et al., 2003). Beneficial, although not statistically significant, effects of the UPC curriculum were also observed for other substance use behaviors, including binge drinking.

We also found the intervention had beneficial effects on adolescents’ perceptions of friends’ use of alcohol, tobacco, and other drugs. These beliefs have been repeatedly linked to initiation of and increased substance use, although this influence is likely bi-directional (Branstetter et al., 2011; Farley & Kim-Spoon, 2015; Simons-Morton, 2007).

UPC training and the school interventions also contributed to significant increases in students’ awareness of appropriately written and communicated school policies related to prohibiting tobacco and alcohol use on school campuses that have also been associated with reductions in substance use (Evans-Whipp et al., 2004). In addition, students in the intervention group reported increases in bonding with their schools, another factor that is protective against substance use (Oelsner et al., 2011).

Finally, we note the significant effects found for reductions in problems related to substance use as well as problems attributable to any reason, both of which are compelling findings to reference in future efforts to persuade schools in Peru and elsewhere to implement the substance use prevention strategies specified in the UPC school training.

Limitations

Findings of this study should be considered in light of potential limitations. Secondary schools that participated in our study may not be representative of all secondary schools in Peru or other Latin American countries. Our data collection period was truncated by one year (out of a planned three) because of the COVID pandemic. As a result, the number of biannual data collection points decreased from six – one pretest and five post-tests – to four, thus attenuating power and our ability to detect longer term effects. Therefore, we may have been unable to detect the full cumulative impact of the multi-pronged prevention strategies, which were originally expected to continue through 2020. We also note that students’ exposure to Unplugged was similar in the intervention and control schools. We thus suspect that the reach and penetration of Unplugged may well have limited our ability to detect significant differences in the use of substances other than tobacco by students in the two groups. The presence of Unplugged may also have constrained our ability to test the effects of UPC training on the development and actions of the prevention leadership action teams, as well as the promotion of substance use-related school policies and a positive school climate. Finally, we note that our inability to detect program effects for the recent (i.e., 30-day) use of marijuana and other drugs may well have been a function of their low prevalence among the targeted age groups and a consequent lack of sufficient statistical power.

That said, this study represents a fairly large, controlled trial that comprised 28 schools, almost all of which participated in all four rounds of data collection. The average student survey response rate (> 85%) was also uniformly high.

Conclusions

Our evaluation of the school curriculum of the UPC series provides initial evidence that UPC training may be an effective means of disseminating prevention science concepts, skills, and practices designed to support the implementation of evidence-based interventions in developing countries. However, we note that the various modules that constitute the UPC series are lengthy and require a major commitment of time and resources on the part of interested departments or ministries that include substance use prevention in their mission. They also require that the modules be taught as originally designed, with the modules designed for prevention coordinators (i.e., mid-level agency staff) preceding those for practitioners providing direct services. Because the modules depend on highly interactive teaching strategies, we are concerned that any move to truncate the length of the modules or to move them to an on-line modality of administration may attenuate the effects noted here.