Evaluation of the effectiveness of a school-based cannabis prevention program

Drug and Alcohol Dependence 132 (2013) 257–264

Contents lists available at SciVerse ScienceDirect

Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep

Evaluation of the effectiveness of a school-based cannabis prevention program Carles Ariza a,b,c,∗ , Anna Pérez a,b,c , Francesca Sánchez-Martínez a,b,c , Marta Diéguez a , Albert Espelt a,b,c,d , M. Isabel Pasarín a,b,c , Josep M. Suelves e , Rafael De la Torre f,g,h , Manuel Nebot a,b,c,1 a

Agència de Salut Pública de Barcelona, Barcelona, Spain Biomedical Research Center Network for Epidemiology and Public Health (CIBERESP), Spain Biomedical Research Institute St. Pau (IBB-St.Pau), Barcelona, Spain d Department of Psychobiology and Methodology of Health Sciences, Autonomous University of Barcelona, Spain e Agència de Salut Pública de Catalunya, Generalitat de Catalunya, Barcelona, Spain f IMIM (Hospital del Mar Research Institute), Barcelona, Spain g Department of Experimental and Health Sciences (DCEXS), Universitat Pompeu Fabra, Barcelona, Spain h CIBER de Fisiopatología de la Obesidad y Nutrición (CB06/03), CIBEROBN, Santiago de Compostela, Spain b c

a r t i c l e

i n f o

Article history: Received 17 October 2012 Received in revised form 8 February 2013 Accepted 10 February 2013 Available online 9 March 2013 This paper is dedicated to Manuel Nebot. Keywords: Cannabis Preventive program Evaluation Effectiveness School-based Process evaluation

a b s t r a c t Background: The effectiveness of a cannabis prevention program in high school students was assessed. Methods: A quasi-experimental study was designed to evaluate the effectiveness of an intervention implemented in an intervention group (IG) with 39 schools compared with a control group (CG) of 47 schools not exposed to the intervention. Of 224 secondary schools in Barcelona, 86 were assessed in the 2005–2006 school year through a personal questionnaire administered at baseline and 15 months after the intervention. Participants consisted of 4848 ninth graders (14–15 year-olds), 2803 assigned to the IG and 2043 to the CG, according to the type and size of the school and the socioeconomic status of the school’s neighborhood. The intervention consisted of a school-based cannabis prevention program (xkpts.com), with four sessions and 16 activities, implemented over 6–10 h, with materials for parents and web-based student involvement. Last-month cannabis use was assessed at baseline and at 15 months’ follow-up. Process evaluation indicators were assessed. Results: At 15 months follow-up, 8.2% of boys and 8.3% of girls in the IG became last-month cannabis users versus 11.8% of boys and 11.6% of girls in the CG. These differences were statistically significant (p = 0.003), representing a 29% reduction in last-month cannabis users in the IG compared with the CG. The incidence of last-month cannabis use was lowest in classrooms that adhered to the program protocol. Conclusions: The xkpts.com program was effective in preventing progression to last-month cannabis use. Effectiveness was higher in classrooms that adhered closely to the protocol. © 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Cannabis is the most widely consumed illegal drug in Europe (Swedish Council for Information on Alcohol and other Drugs (CAN), 2009). Early onset of consumption during adolescence has been related to the presence of several learning problems, low selfesteem and depression (Coffey et al., 2000; Von Sydow et al., 2002; Macleod et al., 2004; Fontes et al., 2011) and increases the risk of cannabis addiction (De Graaf et al., 2010). In addition, a bi-directional causal association between cannabis use and

∗ Corresponding author at: Evaluation and Intervention Methods Service, Agència de Salut Pública de Barcelona (Public Health Agency, Barcelona), Pl. Lesseps, 1, 08023 Barcelona, Spain. Tel.: +34 93 202 77 43; fax: +34 93 292 14 43. E-mail address: [email protected] (C. Ariza). 1 Deceased. 0376-8716/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.drugalcdep.2013.02.012

vulnerability to psychosis has been demonstrated (Kuepper et al., 2011; Griffith-Lendering et al., 2012). Prevention programs to reduce the number of young persons progressing from ever use to regular cannabis use are clearly needed. Drug dependency preventive interventions, including those related to cannabis use, are highly disseminated in Europe, but most are not supported by scientific evidence of effectiveness (European Monitoring Centre for Drugs and Drug Addiction (EMCDDA), 2009). The development of prevention strategies based on evidence is essential to improve their community effectiveness and to avoid the choice of ineffective and sometimes harmful interventions (Faggiano, 2010a). School-based prevention of cannabis use may be effective in high schools when preventive programs include elements from different theoretical models as opposed to programs based solely on the social influence model (Porath-Waller et al., 2010). In this study the intervention program was designed following the principles of

258

C. Ariza et al. / Drug and Alcohol Dependence 132 (2013) 257–264

Fig. 1. Flow-chart of the matching process in the follow-up.

other recognized interventions that adopt the “life skills training” model (Botvin et al., 1990; Sussman et al., 2002; Ellickson et al., 2003). Adding family-focused interventions to community-based or school-based interventions increases their effectiveness (Kumpfer et al., 2002). Consequently, this study included a family component to allow interaction between parents and their children. The components of effective parent and family programs include addressing strategies to improve family relations, communication and parental monitoring (Kumpfer and Alvarado, 2003). In Barcelona, experimental cannabis use doubled in 5 years, progressing from 17.4% among 14–15-year-olds in 2000 (Nebot et al., 2006; Guxens et al., 2007) to 35.1% among students of the same age in 2005 (Morales et al., 2008; Pérez et al., 2009). Consequently, a cannabis prevention program was designed and implemented in the 2005–06 school year. The main objective of this study was to evaluate the effectiveness of this preventive intervention in a sample of 14–15-year-old students. 2. Methods 2.1. Design and sample selection A quasi-experimental study was designed to evaluate the effectiveness of an intervention implemented in the intervention group (IG) compared with a control group (CG) not exposed to the intervention. As inclusion criteria, participating schools had to have confirmed their acceptance and have previously applied the drug dependency preventive program “Decideix” (Calafat et al., 1995) in the third year of high school (14–15-year-old age group, ninth grade), thus ensuring that all participating schools had a similar preventive level at baseline. Of 224 high schools in Barcelona, 93 met these requirements. The assignment of schools to two groups took into account the type of school (public versus subsidized/private), the school’s size (number of students in the third year of high school) and the socioeconomic status of the school’s neighborhood. According to this stratified sampling, the 93 schools were randomly assigned to one of the two groups: 41 to the IG and 52 to the CG. Seven schools, two in the IG and five in the CG, refused to participate in the project, because they objected to the conditions related to the evaluation (test and re-test 1 year later). Thus, 39 schools (117 classrooms with 3024 students) were included in the IG and the remaining 47 (97 classrooms with 2259 students) in the CG. The students were aged 14–15 years old. Some students were absent in the post-test or left the school between the two surveys; thus, these questionnaires could not be matched between pre-test and post-test. As a result, the post-test questionnaire was matched with the pre-test in 1863 of 2805 students (66.2%) in the IG and in 1328 of 2043 (65.0%) in the CG. Overall, attrition at the end of follow-up was 33.8% in the IG and 35.0% in the CG (Fig. 1) and was similar between the two groups.

between groups, the baseline questionnaire was administered between January and March, 2006, and the follow-up questionnaire between April and May, 2007, 15 months after the intervention ended (Fig. 2). The intervention had been implemented in April and May, 2006. A confidential, alpha-numeric code based on students’ date of birth and the initials of their parents’ names allowed baseline questionnaires to be linked to follow-up questionnaires. Finally, a self-reported questionnaire, addressed to teachers participating in the intervention, was used for process evaluation. The completeness (number of activities implemented in the classroom) and fidelity to the program (implementation of activities proposed in the protocol) were collected. 2.3. Variables 2.3.1. Dependent variable (outcome criterion). The dependent variable, the cumulative incidence rate (CIR) was defined as the change in reported cannabis use between baseline and the follow-up at 15 months. This variable was constructed to identify “non users” (those who had never tried cannabis), or “lifetime users” (those who had used cannabis at least once but not in the last month) at baseline that progressed to “use in the last 30 days” (last-month users or regular users) in the follow-up questionnaire. 2.3.2. Explanatory individual variables (predictors). Individual information on sociodemographic variables, family situation, self-perceived academic performance, and weekly personal allowance were collected. Information on the family situation consisted of living in two-parent households or other situations. Students indicated their perceived relative position regarding academic performance (high, medium or low). Their weekly allowance was categorized as 0 D , less than 10 D , 10–30 D and more than 30 D . The students were also asked about substance use such as tobacco and alcohol. Occasional smokers were those who reported smoking cigarettes at least once a month but not every week. Regular smokers were those who reported smoking at least once a week. Risky alcohol consumption was defined as having been drunk at least once. Questions on leisure time concerned going out to bars or discotheques. “skipping class” (never, once or more times) was studied as one of several antisocial behaviors (Nebot et al., 2006). Among the psychosocial variables, “self-efficacy” was defined as the ability to refuse an offer to consume cannabis products. Students were also asked about ease of access to cannabis and about risk perception of its use. The role of expectancies about the effects of cannabis use was measured through six items obtained by factorial

2.2. Procedure Data were obtained through a self-reported written questionnaire, which reliability and validity had previously been explored (Moncada and Pérez, 2002) and adapted to the study of cannabis use (Nebot et al., 2006). The questionnaire was administered during 1 h of class time by personnel from the Public Health Agency of Barcelona without the participation of teachers. In the CG, baseline data were obtained between January and March, 2005, while the follow-up questionnaire was administered between April and May, 2006. In the IG, to avoid contamination

Fig. 2. Design of the cannabis prevention program evaluation study.

C. Ariza et al. / Drug and Alcohol Dependence 132 (2013) 257–264

259

Table 1 Intervention: main activities of the “xkpts” program. Level of intervention

Session number

Contents

Techniques and materials

Number of activities

Individual (classroom)

1

Concepts of drug dependencies Taking decisions

6

Individual (classroom) Family

2a 2b

Individual (home)

2c

Individual (classroom)

3

Situations of cannabis use. Social influences Information about cannabis Parents advised to speak about cannabis with their sons and daughters DVD continuation. Information about cannabis Stories of cannabis use among students’ peers How to deal with cannabis offers. Learning to refuse

Presentations. Discussion Sheets DVD: “xkpts.com” Cannabis: Let’s speak at home (family guide) Form to complete

Individual (classroom)

4

Consequences of drug abuse Ways to proceed

Website: http://www.xkpts.com Life skill training Role playing Discussion Sheets

3 a

a

3 4

ACCEPTABLE Implementation: completion of 8 out of 16 activities. QUALIFIED Implementation: performance of the specific cannabis sessions (2a, 3) with fidelity to the protocol. a There was no quantitative control of these activities.

analysis from the short version of the “Marijuana Effects Expectancy Questionnaire” (Schafer and Brown, 1991; Aarons et al., 2001). Adolescents were asked about the expected consequences of cannabis on the intellect, behavior, relaxation, social and sexual relations, cognitive and perceptive capacity, and health. A summary score was calculated (0–6, depending on the number of favorable expectancies) and was treated as a continuous variable in the analysis. 2.3.3. Explanatory contextual variables. The main contextual independent variable was “being exposed to the cannabis intervention program” (yes or no). Moreover, in the IG, the number of activities implemented was measured. Classrooms that completed more than eight activities were classified as belonging to the “acceptable” IG and those that used the techniques (DVD and role playing) recommended in the protocol were included in the “qualified” IG. Other contextual variables were also taken into account. Schools were categorized into public, private or subsidized. As an approximation of socioeconomic position, the Family Economic Capacity Index (FECI), developed and validated in 1996 in Barcelona, was used (Ventura et al., 1999). This index is an ecologic indicator that reflects the wealth of the school’s neighborhood or the pupil’s neighborhood of residence. The FECI was categorized as high, medium or low. 2.4. Intervention The intervention consisted of a school-based cannabis prevention program named “xkpts.com” (reader: “perquèpetes.com” and translation “why joints?”), addressed to 14–16-year-olds. This program adopts a universal approach and was recommended for use in high school students in the ninth grade (third grade of compulsory secondary education in the Spanish educational system). The individual responsible for its implementation was the teacher in charge of each class. The Public Health Agency of Barcelona funded materials and provided training to teachers and technical support through community health teams. The rationale for the program was the Attitudes-Social influences-self-Efficacy (ASE) model (De Vries et al., 1995). Briefly, this model holds that learning healthy or unhealthy behavior is influenced by the intention to adopt a behavior in the future, favorable attitudes or expectancies, social influences (modeling, social pressure, subjective norms) and self-efficacy. All these variables were considered as intermediate variables in our intervention and the distinct sessions of the program specifically addressed each of them. The program included four sessions (Table 1), with 16 activities, to be implemented in 6–10 class hours. A detailed description of the intervention and its evaluation process has previously been published (Sánchez-Martínez et al., 2010). The teacher’s material included a guide with instructions on how to implement the program (Ariza et al., 2006a,b) and a DVD with a story that contextualized cannabis use in the student environment. Families received a guide at home (Ariza et al., 2006a,b) with materials for speaking about this topic with their sons and daughters. In addition, teachers invited students to enter the web http://www.xkpts.com to find other complementary resources related to the topic (Table 1). Participating teachers had to complete a process evaluation form. A 6-h training course was offered to new implementers in a district center, outside the school. The main contents of the course were knowledge of the health risks of cannabis and skills training on how to recognize peer pressure to use cannabis and how to refuse it among teenagers. The “xkpts.com” program was evaluated according to the quality criteria of the “Exchange on Drug Demand Reduction Action” (EDDRA) of the European Union (European Monitoring Centre for Drugs and Drug Addiction, EMCDDA) database and was added to this list in 2011 (EMCDDA, 2011).

2.5. Analysis A bivariate analysis was carried out to study differences between the IG and the CG. This analysis focused on matched or unmatched individuals and consumption and contextual variables in sociodemographic and substance use characteristics associated with the follow-up status. Percentages (for qualitative variables) were compared using the chi-square test, while means (for quantitative variables) were compared using Student’s t-test. The CIR of last month use was calculated by the degree of intervention (CG, total IG, acceptable IG and qualified IG). To determine the association between the CIR of last-month use and the degree of implementation, a logistic regression was fitted and odds ratios (ORs) and 95% confidence intervals (CI) were obtained. Finally, the CIR of last-month use was calculated for each explanatory variable or predictor (individual and contextual). The variables related to the CIR of last-month use were included in the multivariate analysis (as possible confounders). Three multilevel logistic regression models were fitted to determine the association between the intervention and the CIR of last month use. Multilevel regression models allowed variance among students within the same school and variance among schools to be taken into account. In the first model (model A), intervention exposure and sociodemographic variables (age and sex, as individual variables, and the type of school and FECI, as contextual variables) were introduced, together with the four mediating variables of the rationale (expectancies; risk perceptions related to cannabis; accessibility to the substance; and self-efficacy). In the second and third model (models B and C), regular smoking and having been drunk, as common predictors of cannabis progression, were added to the first model. To minimize the differences observed at baseline, in the multivariate analysis all models were adjusted by the variables included in each model. The analysis was carried out using the SPSS statistical package, version 13, and HLM 7.0.

2.6. Ethical issues During the study, the national and international guidelines were followed (codes of professional ethics, the Helsinki Declaration of 1964 and subsequent revisions). Additionally, the Spanish law on data confidentiality was observed (Law 15/1999 of 13 December on Personal Data Protection). School staff informed the teenagers’ parents of the aims of the study and obtained their collaboration. All the information was protected by using standard data management and information storage procedures.

3. Results 3.1. Attrition and baseline equivalence Table 2 shows the main socio-demographic characteristics in the IG compared with the CG, as well as the total sample with losses to follow-up (dropouts, non-completers, etc.). Because this study analyzed progression from non-use or lifetime use to last-month cannabis use, 429 (8.8%) last-month cannabis users at baseline were removed [(217 (7.7%) in the IG and 212 (10.4%) in the CG)]. Consequently, 2586 (92.3%) students in the IG and 1831 (89.6%) in the CG were studied for differences in progression of use.

260

Table 2 Sociodemographic characteristics of matched and unmatched samples. Variables

Intervention

Total (a)

Control

Matched (b)

Unmatched (c)

p

Total (d)

Matched (e)

Unmatched (f)

Total I vs total C p (a) vs (d)

Matched I vs matched C p (b) vs (e)

0.000

0.000

p

%

N

%

N

%

N

%

N

%

N

%

14.44

(0.625)

14.29

(0.537)

14.60

(0.727)

14.55

(0.67)

14.43

(0.592)

14.77

(0.747)

1361 1441 2802

48.60 51.40

947 916 1863

50.90 49.10

414 525 939

41.90 58.00

0.010

1006 1036 2042

49.20 50.70

678 650 1328

51.10 48.90

328 386 714

45.90 54.10

0.029

0.634

0.901

653 1415 359 2427 Self-perceived academic performance 408 Low 1639 Middle High 726 2773 Family situation 2172 Two-parent home 582 Single-parent home 49 Other 2803 Weekly pocket money (D ) 1170 0 1126 <10 10–30 435 <30 56 2787 Cannabis use 1996 Never 590 Tried 217 Regular 2803 Tobacco consumption 2187 Non-smoker Occasional (less than weekly) 219 373 Regular (weekly or daily) 2779 Ever drunk 1920 No 857 Yes 2777 Socioeconomic level (FECIa ) of the school 1052 Low Middle 822 High 929 2803 School type 2057 Private/subsidized Public 746 Total 2803

23.30 50.50 12.80

385 974 248 1607

24.00 60.60 15.40

268 441 111 820

32.70 53.80 13.50

0.000

614 955 205 1774

30.10 46.70 10.00

355 661 130 1146

31.00 57.70 11.30

259 294 75 628

41.20 46.80 11.90

0.000

0.000

0.000

14.60 58.50 25.90

164 1106 572 1842

8.90 60.00 31.10

244 533 154 931

26.20 57.30 16.50

0.000

323 1151 538 2012

15.80 26.30 56.30

130 752 428 1310

9.90 57.40 32.70

193 399 110 702

27.50 56.80 15.70

0.000

0.328

0.303

77.50 20.80 1.70

1523 317 23 1863

81.70 17.00 1.20

649 265 26 940

69.00 28.20 2.80

0.000

1560 441 21 2022

76.40 21.60 1.00

1081 227 12 1320

81.90 17.20 0.90

479 214 9 702

68.20 30.50 1.3

0.000

0.095

0.683

41.70 40.20 15.50 2.00

818 776 241 18 1853

44.10 41.90 13.00 1.00

352 350 194 38 934

37.70 37.50 20.80 4.10

0.000

792 885 298 50 2025

38.80 43.30 14.60 2.40

566 576 160 16 1318

42.90 43.70 12.10 1.20

226 309 138 34 707

32.00 43.70 19.50 4.80

0.000

0.067

0.629

71.20 21.00 7.70

1441 315 107 1863

77.30 16.90 5.70

555 275 110 940

59.00 29.30 11.70

0.000

1277 554 212 2043

62.50 27.10 10.40

929 310 89 1328

70.00 23.30 6.70

348 244 123 715

48.70 34.10 17.20

0.000

0.000

0.000

78.00 7.80 13.30

1537 143 166 1846

83.30 7.70 9.00

650 76 207 933

69.70 8.10 22.20

0.000

1457 200 372 2029

71.30 9.80 18.20

1022 133 161 1316

77.70 10.10 12.20

435 67 211 713

61.00 9.40 29.60

0.000

0.000

0.000

68.50 30.60

1405 445 1850

75.90 24.10

515 412 927

55.60 44.40

0.000

1327 687 2014

65.00 34.10

951 360 1311

72.50 27.50

376 327 703

53.50 46.50

0.000

0.017

0.006

37.5 29.3 33.1

675 497 691 1863

36.20 26.70 37.10

377 325 238 940

40.01 34.60 25.30

0.000

468 914 661 2043

22.90 44.70 32.40

285 624 419 1328

21.50 47.00 31.60

183 290 242 715

25.6 40.6 33.8

0.014

0.000

0.000

73.4 26.6 100.0

1480 383 1863

79.40 20.60 100.0

577 363 940

61.40 38.60 100.0

0.000

1309 734 2043

64.10 35.90 100.0

917 411 1328

69.10 30.90 100.0

392 323 715

55 45 100.0

0.000

0.000

0.000

Socioeconomic level (FECIa ) Low Middle High

a

Family economic capacity index, see Section 2.

C. Ariza et al. / Drug and Alcohol Dependence 132 (2013) 257–264

N Age (mean age) Sex Male Female

C. Ariza et al. / Drug and Alcohol Dependence 132 (2013) 257–264

There were statistically significant differences between the IG and the CG at baseline, as IG participants were younger (6 months), more lived in neighborhoods with a higher socioeconomic level (FECI), and more attended private/subsidized schools. There were also differences in cannabis use (less lifetime and last month users), smoking (fewer occasional and regular smokers), and alcohol (fewer people “ever drunk”). When matched samples in the IG and CG were compared, attrition analysis showed the same significant differences between groups as those previously described (Table 2). 3.2. Cannabis use and effectiveness evaluation Table 3 summarizes progression from non-cannabis use or lifetime use at baseline to last-month cannabis use at 15 months in the two groups by the degree of implementation of the intervention and sex. There were no differences in sex between the two groups. In the IG, 8.2% of boys and 8.3% of girls were new lastmonth cannabis users at 15 months versus 11.8% of boys and 11.6% of girls in the CG. The difference in CIR between the groups was 3.4%, which was statistically significant (p = 0.003). When progression in the CG was compared with that in students from the “acceptable IG” subgroup, the difference in CIR was 4.0%; this difference increased to 4.4% in students from the “qualified IG” subgroup. In all cases, differences in CIR among these subgroups were also statistically significant (p = 0.002 and p = 0.018, respectively). The intervention produced a 29% reduction in last-month cannabis users (percentage of possible users avoided) when the IG was considered as a whole and reductions of 34% and 36%, when the “acceptable IG” or “qualified IG” subgroups, respectively, were analyzed. When the effect of the intervention, as a contextual variable, was examined on progression to last-month use in the multilevel analysis, the OR was 1.40 [95% CI: 1.03–1.91] for the IG as a whole; 1.51 [95% CI: 1.01–2.27] for the “acceptable IG” subgroup and 1.67 [95% CI: 1.04–2.69] for the “qualified IG” subgroup. The model was adjusted for four sociodemographic variables, age and sex, as individual variables, and type of school and socioeconomic status (FECI) (see Section 2), as contextual variables. Table 4 shows the multilevel analysis of variables included in the first model (model A: sociodemographic and mediating variables). The main statistically significant factors associated with progression were “favorable expectancies of cannabis use”, “intervention” (belonging to the CG), “type of school” (public), “cannabis risk perception” (non- or little dangerous), “low self-efficacy to cope with cannabis offers” and “easy access to cannabis.” When other behavioral variables commonly related to cannabis use were added (models B and C), only “cannabis risk perception”, “low self-efficacy to cope with cannabis offers” and “easy access to cannabis” retained significance. In addition, sex (being male) and the four behavioral variables (skipping class, going to discotheques, having been drunk and regular smoking) also showed a statistically significant effect. The difference in models B and C lay in smoking consumption and its influence on the effect of the intervention program. When this variable took in account all the students who progressed to regular cannabis use, independently of the kind of smoking consumption at baseline (model B), the effect of the intervention disappeared [OR = 1.29 (0.64–2.60)], but when only students who were non-smokers at baseline were considered (model C), the effect of the intervention was mantained [OR = 1.53 (1.02–2.31)]. 4. Discussion 4.1. Principal findings The “xkpts.com”, school-based cannabis prevention program, designed to deal specifically with the sudden increase of cannabis

261

use among Barcelona students between 2000 and 2005 (Nebot et al., 2006; Morales et al., 2008) is presented. We show that a reduction of approximately 37% in last-month cannabis users would be achieved with the program. This success rate would be explained by the family and web-based peer involvement together with the full implementation of all the other components of the program. The results show that, at 15 months post-intervention, there were 29% fewer last-month cannabis users in the IG compared with the CG. This rate rose to 34% and 36%, depending on the completeness of the implementation of the intervention (“acceptable IG” or “qualified IG” subgroups, respectively). In the multivariate analysis, after a multilevel analysis to control for the cluster effect, the factors associated with cannabis use at 15 months’ follow-up were “attending a public school” and “being unexposed to the intervention”, together with the main mediating variables of “having favorable expectancies of cannabis use”, “perceiving cannabis as non- or little dangerous”, “showing low self-efficacy to cope with cannabis offers” and “easy access to cannabis.” The positive results found by this program in last-month cannabis use are consistent with those of previous studies with different follow-up periods. In fact, our program integrated the different rationales used in some of these studies (social influence perspective, life skills training), as well as the use of Internet resources, in a single framework. Newton et al. (2009) demonstrated that an intervention program on marijuana use had a significant effect at 6 months follow-up, Botvin et al. (1984) and Ellickson et al. (2003) at 1 year, Sussman et al. (2002) at 2 years, Furr-Holden et al. (2004) at 5 years and Botvin et al. (1990) after 6 years. Significant differences of 3–4% in last-month cannabis CIR and reductions of 20–30% in new last-month users, similar to those in the current study, were also identified in the Gatehouse Project (Bond et al., 2004) and in the Lions-Quest Skills for Adolescents (Eisen et al., 2003) after the second year of the program. The EU-DAP Trial, developed in seven European countries, showed a 24% reduction in last-month cannabis users after the intervention (Faggiano et al., 2010b). 4.2. Strengths and weaknesses School-based programs without additional components are usually ineffective (Botvin, 1999; Flay, 2000; Lloyd et al., 2000). One of the strengths of the “xkpts.com” program is the family component. Parents were incorporated into the intervention through a guide that they received at home at the middle of the school implementation. The guide may have prompted them to speak about cannabis with their sons and daughters and to give them possible strategies for finding solutions. The program asked parents to provide feedback to the research team about their views and outcomes. Unexpectedly, 70% of families (data published previously, Sánchez-Martínez et al., 2010) provided this feedback. Several studies have reported that the effectiveness of interactive programs increased when they also included this family component, particularly if the programs produced cognitive, affective and behavioral changes in the ongoing family dynamics and environment (Kumpfer and Alvarado, 2003; Longshore et al., 2007; Faggiano et al., 2008). The influence of interactive components to prevent drug dependency on the effectiveness of these programs has been well established (Tobler et al., 1999; Cuijpers, 2002; McGrath et al., 2006). Tobler et al. (1999) reported that the best results in the incidence of marijuana use were obtained when interactivity included the development of social competencies. A strength of the “xkpts.com” intervention program was an accurate process evaluation with measurement of the degree of implementation not only in each school, but also within each class. Consequently, the

262

C. Ariza et al. / Drug and Alcohol Dependence 132 (2013) 257–264

Table 3 Cumulative incidence rate (CIR) of last-month use by degree of implementation of the intervention and sex. Cumulative incidence rate (CIR) of last month use

Control group Whole Intervention group Acceptable intervention group Qualifiedd intervention group

Boys

Girls

a

a

Difference in AIR

Ratio of incidencesb

p

OR (95% CI)c

3.4 4 4.4

29.0% 34.1% 36.7%

0.003 0.002 0.018

1.408 (1.035–1.914) 1.519 (1.015–2.276) 1.678 (1.046–2.694)

Total

N

%

N

%

Na

%

72/610 71/862 27/372 18/246

11.8 8.2 7.3 7.3

73/629 74/894 36/451 21/288

11.6 8.3 8 7.3

145/1239 145/1756 63/823 39/534

11.7 8.3 7.7 7.3

ACCEPTABLE implementation: completion of 8 out of 16 activities. QUALIFIED implementation: performance of the specific cannabis sessions (2a, 3) with fidelity to the protocol. a Number of new last-month cannabis users (numerator) among non-users or lifetime users at baseline (denominator). b Ratio of incidences (percentage of possible last-month users avoided). c Multilevel analysis, adjusted by sociodemographic variables: age and sex (individual variables) and type of school and FECI (see Section 2) (contextual variables). d See Section 2. Table 4 Multilevel analysis of factors associated with progression to last-month cannabis use at follow-up. Model B: variables Model A: Model C: variables included in model sociodemographic included in model B, but A + other behavioral only considering variables + mediating variables (rationale) variables non-smokers at baseline Non-users or lifetime users at baseline that progressed to last-month users at follow-up (after 15 months). Cumulated incidence rate (CIR) n Individual variables 14.42 Age (mean age) (SD) Gender 147 Female 142 Male Risk perception of cannabis 44 Very dangerous 245 Non- or a little dangerous Expectancies of cannabis use 2.82 Access to cannabis Not easy 62 219 Easy Self-efficacy to cope with cannabis offers 192 Yes 81 No Smoking 159 No Occasional 59 70 Regular Having been drunk 136 Never 153 Once or more times Going to discotheques 73 Never 216 Once or more times Skipping class 195 Never 94 Once or more times

%

OR (95% CI)a

OR (95% CI)a

OR (95% CI)a

(0.619)

1.187 (0.947–1.488)

0.949 (0.741–1.215)

0.957 (0.747–1.225)

9.7 9.7

1 1.125 (0.859–1.473)

1 1.568 (1.165–2.111)

1 1.567 (1.164–2.110)

5.1 11.6 (1.44)

1 1.807 (1.351–2.611) 1.113 (1.011–1.225)

1 1.555 (1.046–2.312) 1.062 (0.958–1.178)

1 1.550 (1.042–2.306) 1.060 (0.956–1.175)

4.5 14.2

1 3.155 (2.298–4.331)

1 2.102 (1.494–2.957)

1 2.103 (1.494–2.959)

8.2 17.2

1 2.276 (1.674–3.094)

1 1.665 (1.195–2.321)

1 1.674 (1.200–2.333)

6.3 25.4 33.7

1 2.486 (1.664–3.714) 3.227 (2.116–4.923)

1 3.219 (1.841–5.628) 3.506 (1.954–6.290)

5.9 22.8

1 2.249 (1.634–3.096)

1 2.241 (1.628–3.084)

5.1 14.0

1 1.739 (1.243–2.434)

1 1.746 (1.247–2.443)

8.0 17.2

1 1.399 (1.006–1.945)

1 1.412 (1.015–1.964)

1 1.405 (1.013–1.951)

1 1.360 (0.961–1.887)

1 1.535 (1.019–2.313)

1 1.062 (0.702–1.605) 1.462 (0.973–2.196)

1 0.968 (0.624–1.502) 1.356 (0.882–2.084)

1 0.714 (0.473–1.079) 1.746 (0.485–1.148)

1 1.485 (1.033–2.133)

1 1.357 (0.917–2.007)

1 1.352 (0.914–1.999)

Contextual variables Being exposed to the cannabis prevention program (xkpts.com) 145 8.3 Intervention group 145 11.7 Control group Socioeconomic status (FECIb ) 72 8.0 High 99 9.4 Middle 119 11.4 Low Type of school 201 8.9 Private/subsidized 89 12.1 Public a b

Each variable was adjusted by the included variables in the model. See Section 2.

number of classes that developed skills training sessions according to the protocol could be identified. Thus, the schools with the best implementation (the qualified IG subgroup) achieved the best results in the incidence of last-month cannabis use (7.3% versus

11.7% in the CG; p = 0.018). As other authors have stated, programs are more effective in reducing cannabis use when high fidelity of implementation is achieved (Dusenbury et al., 2003; Coggans et al., 2003).

C. Ariza et al. / Drug and Alcohol Dependence 132 (2013) 257–264

One of the limitations of this study concerns its internal validity. In Barcelona, pre-existing drug dependency programs made randomization very difficult. Consequently, a quasi-experimental design was chosen to minimize contamination, including schools that had previously been exposed to the same pre-existing drug dependency program. Furthermore, new components of a specific cannabis intervention were added to the IG. In addition, schools were allocated to groups according to different socioeconomic strata and type of school. As described by other authors (Schofield et al., 2003; Crone et al., 2003) more schools (and especially larger schools) refused to participate in the CG than in the IG, causing differences in behavioral and sociodemographic data among groups at baseline. To correct for these confounding factors, progression to last-month cannabis use in the multivariate analysis was adjusted for all these variables. A potential problem concerning internal and external validity is attrition. The average attrition rate was 34%, and was similar among groups. Participants remaining in the study, as opposed to those who withdrew, can be characterized as being slightly younger, attending a private/subsidized school, living in neighborhoods with middle-high socioeconomic status and more frequently in two-parent families, having middle or high self-perceived academic performance and less pocket money, being less likely to be a regular smoker or regular cannabis user, and being less likely to have been drunk. Nevertheless, these variables, except sex, school type and smoking and drunkenness, did not influence the effect of the intervention when progression to last-month cannabis use at the end of the study was adjusted for these variables in the logistic regression. In addition, analysis of losses showed that the bias was similar in the two groups (IG and CG) (Table 2) and therefore its impact on the estimation of the effects was negligible. Another limitation emerged in the multilevel analysis when other behavioral variables strongly correlated with cannabis use were introduced, such as “being a regular smoker” or “having been drunk” (models B and C, Table 4). In the case of model B, when we considered students who progressed to regular cannabis use, independently of the kind of their smoking consumption, the protective effect of the intervention was lost. It is well known that, with a given power of a study, the study sample needs to be enlarged in order to control for this effect (Campbell et al., 2001). In contrast, the strong association of variables such as “smoking regularly” and “had been drunk” led to a narrowness of the sample, a decrease of the global statistical power, and the loss of the intervention effect. However, when we considered only students who progressed to regular cannabis use and were also non-smokers at baseline (model C, Table 4), the protective effect of the intervention was maintained. This finding confirmed that a universal approach, such as that adopted by the “xkpts” program is effective in maintaining persons who did not use drugs other than cannabis at baseline as non-users. In addition, among students who were users at baseline, whether of cannabis or other drugs, more selective strategies than those used in the “xkpts” program are needed. In summary, the “xkpts” cannabis preventive program was effective in reducing the number of secondary school students progressing to regular cannabis use. This effect was enhanced by greater fidelity to the protocol when the program was implemented. 4.3. Implications for further study The results of this study show that the “xkpts.com” program is an effective strategy to deal with the sudden increase of cannabis

263

use among secondary school students in Barcelona. As occurred in California several years ago, the increasing social acceptance of cannabis required a specific preventive universal program that introduced social issues concerning the substance and credible messages to youth (Lafferty, 1998). To improve the program’s dissemination and effectiveness, its incomplete implementation should be addressed. A possible reason for this incompleteness may have been that the teachers felt uncomfortable with certain activities such as the role plays included in the protocol (Botvin et al., 1990). Accordingly, more extensive training could be used to convince teachers of the merits of these kinds of strategies and provide them with sufficient confidence in their implementation. Finally, more studies are needed to evaluate the effect of each of the program’s components in relation to the overall effect of the intervention. Determining what part of the effectiveness is due to each part of the program (school-based, family-based, peer-based) would be highly useful. Role of funding source Funding for this study was provided by a National Plan for Drugs (Spanish Ministry of Health) grant (SCO/3246/2004) for expenses related to the evaluation of the program and by a “Viure i Conviure” Foundation (Caixa Catalunya Charity Foundation) grant for expenses related to the intervention program. Contributors C. Ariza, A. Pérez and M. Nebot designed the study. F. SánchezMartínez and M. Diéguez participated in the literature searches and the procedure. A. Pérez, F. Sánchez-Martínez and A. Espelt undertook the statistical analysis. C. Ariza wrote the first draft of the manuscript. M.I. Pasarín, J.M. Suelves, R. De la Torre and M. Nebot provided critical input to múltiple drafts of the paper. All authors contributed to and approved the final manuscript. Conflict of interest statement All authors declare that they have no conflicts of interest. Acknowledgements We thank nurses from the Community Health Service of the Public Health Agency of Barcelona and the 86 schools (coordinators, teachers and students) for their important contribution to the study. References Aarons, G.A., Brown, S.A., Stice, E., Coe, M.T., 2001. Psychometric evaluation of the marijuana and stimulant effect expectancy questionnaires for adolescents. Addict. Behav. 26, 219–236. Ariza, C., Pérez, A., Nebot, M., Juárez, O., Rodríguez-Martos, A., 2006a. Cannabis Consumption Preventive Program “xkpts.com” (Protocol, DVD and Teacher’s Guide). Public Health Agency Viure i Conviure Fundation. Caixa Catalunya Charity Foundation, Barcelona. Ariza, C., Rodríguez-Martos, A., Vecino, C., Guitart, A., 2006b. Cannabis: Let’s Speak at Home (Family Guide). Public Health Agency Viure i Conviure Fundation. Caixa Catalunya Charity Foundation, Barcelona. Bond, L., Thomas, L., Coffey, C., Glover, S., Butler, H., Carlin, J.B., Patton, G., 2004. Long-term impact of the Gatehouse Project on cannabis use of 16-year-olds in Australia. J. School Health 74, 23–29. Botvin, G.J., Baker, E., Renick, N.L., Filazzola, A.D., Botvin, E.M., 1984. A cognitive–behavioral approach to substance abuse prevention. Addict. Behav. 9, 137–147. Botvin, G.J., Baker, E., Dusenbury, L., Tortu, S., Botvin, E.M., 1990. Preventing adolescent drug abuse through a multimodal cognitive behavioral approach: results of a 3 year study. J. Consult. Clin. Psychol. 58, 437–446.

264

C. Ariza et al. / Drug and Alcohol Dependence 132 (2013) 257–264

Botvin, G.J., 1999. Prevention in schools. In: Ammerman, R.T., Ott, P. (Eds.), Prevention and Societal Impact of Drug and Alcohol Abuse. Lawrence Erlbaum Associates, Mahwah, NJ. Calafat, A., Amengual, M., Guimerans, C., Rodriguez-Martos, A., Ruiz, R., 1995. “You decide”: 10 years of a school-based prevention program. Adicciones 7, 509–526 (Spanish). Campbell, M.K., Mollison, J., Grimshaw, J.M., 2001. Cluster trials in implementation research: estimation of intracluster correlation coefficients and sample size. Stat. Med. 20, 391–399. Coffey, C., Lynskey, M., Wolfe, R., Patton, G.C., 2000. Initiation and progression of cannabis use in a population-based Australian adolescent longitudinal study. Addiction 95, 1679–1690. Coggans, N., Cheyne, B., McKellar, S., 2003. The Life Skills Training Drug Education Programme: A Review of Research. University of Strathclyde, Scottish Executive Drug Misuse Research Programme, Edinburgh. Crone, M.R., Reijneveld, S.A., Willemsen, M.C., van Leerdam, F.J., Spruijt, R.D., Sing, R.A., 2003. Prevention of smoking in adolescents with lower education: a school based intervention study. J. Epidemiol. Community Health 57, 675–680. Cuijpers, P., 2002. Effective ingredients of school-based drug prevention programs. A systematic review. Addict. Behav. 27, 1009–10023. De Graaf, R., Radovanovic, M., Van Laar, M., Fairman, B., Degenhardt, L., AguilarGaxiola, S., Bruffaerts, R., de Girolamo, G., Fayyad, J., Gureje, O., Haro, J.M., Huang, Y., Kostychenko, S., Leˇıpine, J.P., Matschinger, H., Medina Mora, M.E., Neumark, Y., Ormel, J., Posada-Villa, J., Stein, D.J., Tachimori, H., Wells, J.E., Anthony, J.C., 2010. Early cannabis use and estimated risk of later onset of depression spells: epidemiologic evidence from the population-based World Health Organisation World Mental Health Survey Initiative. Am. J. Epidemiol. 172, 149–159. De Vries, H., Backvier, E., Kok, G., Dijkstra, M., 1995. The impact of social influences in the context of attitude, self efficacy, intention and previous behavior as predictors of smoking onset. J. Appl. Soc. Psychol. 25, 237–257. Dusenbury, L., Brannigan, R., Falco, M., Hansen, W.B., 2003. A review of research on fidelity of implementation: implications for drug abuse prevention in school settings. Health Educ. Res. 18, 237–256. Eisen, M., Zellman, G.L., Murray, D.M., 2003. Evaluating the Lions-Quest “Skills for Adolescence” drug education program. Second-year behavior outcomes. Addict. Behav. 28, 883–897. Ellickson, P.L., McCaffrey, D.F., Ghosh-Dastidar, B., Longshore, D.L., 2003. New inroads in preventing adolescent drug use: results from a large-scale trial of project ALERT in middle schools. Am. J. Public Health 93, 1830–1836. European Monitoring Centre for Drugs and Drug Addiction (EMCDDA), 2009. Annual Report. Lisbon. European Monitoring Center for Drugs and Drug Addiction (EMCDDA), 2011. Exchange on Drug Demand Reduction Action” (EDDRA). “xkpts.com” Program (A cannabis prevention program for teens). Public Health Agency, Barcelona. Available at: http://www.emcdda.europa.eu/modules/ wbs/dsp print project description.cfm?project id=ES 06 (accessed 09.03.12). Faggiano, F., Vigna-Taglianti, F., Versino, E., Zambon, A., Borraccino, A., Lemma, P., 2008. School-based prevention for illicit drugs use: a systematic review. Prev. Med. 46, 385–396. Faggiano, F., 2010a. Prevention of substance abuse: randomised or observational evaluation is absolutely needed. Adicciones 22, 11–14. Faggiano, F., Vigna-Taglianti, F., Burkhart, G., Bohrn, K., Cuomo, L., Gregori, D., Panella, M., Scatigna, M., Siliquini, R., Varona, L., van der Kreeft, P., Vassara, M., Wiborg, G., Galanti, M.R., E.U.-Dap Study Group, 2010b. The effectiveness of a schoolbased substance abuse prevention program: 18 month-follow-up of the EU-Dap cluster randomized controlled trial. Drug Alcohol Depend. 108, 56–64. Flay, B.R., 2000. Approaches to substance use prevention utilizing school curriculum plus social environment change. Addict. Behav. 25, 861–885. Fontes, M.A., Bolla, K.I., Cunha, P.J., Almeida, P.P., Jungerman, F., Laranjeira, R.R., Bressan, R.A., Lacerda, A.L.T., 2011. Cannabis use before age 15 and subsequent executive functioning. Br. J. Psychiatry 198, 442–447. Furr-Holden, C.D.M., Ialongo, N.S., Anthony, J.C., Petras, H., Kellam, S.G., 2004. Developmentally inspired drug prevention: middle school outcomes in a school-based randomized prevention trial. Drug Alcohol Depend. 73, 149–158. Griffith-Lendering, M.F., Wigman, J.T., Prince van Leeuwen, A., Huijbregts, S.C., Huizink, A.C., Ormel, J., Verhulst, F.C., van Os, J., Swaab, H., Vollebergh, W.A., 2012. Cannabis use and vulnerability for psychosis in early adolescence – a TRAILS

study. Addiction (December), http://dx.doi.org/10.1111/add.12050 [Epub ahead of print]. Guxens, M., Nebot, M., Ariza, C., 2007. Age and sex differences in factors associated with the onset of cannabis use: a cohort study. Drug Alcohol Depend. 88, 234–243. Kuepper, R., van Os, J., Lieb, R., Wittchen, H.U., Höfler, M., Henquet, C., 2011. Continued cannabis use and risk of incidence and persistence of psychotic symptoms: 10 year follow-up cohort study. BMJ 342, 738. Kumpfer, K.L., Alvarado, R., Tait, C., Turner, C., 2002. Effectiveness of school-based family and children’s skills training for substance abuse prevention among 6–8year-old rural children. Psychol. Addict. Behav. 16, 565–571. Kumpfer, K.L., Alvarado, R., 2003. Family-strengthening approaches for the prevention of youth problem behaviors. Am. Psychol. 58, 457–465. Lafferty, L., 1998. Marijuana use prevention: the In-DEPTH model program. J. Psychoactive Drugs 30, 205–208. Lloyd, C., Joyce, R., Hurry, J., Asthon, M., 2000. The effectiveness of primary school drug education. Drugs Educ. Prev. Pol. 7, 109–126. Longshore, D., Ellickson, P.L., McCaffrey, D.F., St Clair, P.A., 2007. School-based drug prevention among at-risk adolescents: effects of ALERT plus. Health Educ. Behav. 34, 651–668. Macleod, J., Oakes, R., Copello, A., Crome, I., Egger, M., Hickman, M., Oppenkowski, T., Stokes-Lampard, H., Davey Smith, G., 2004. Psychological and social sequelae of cannabis and other illicit drug use by young people: a systematic review of longitudinal, general population studies. Lancet 363, 1579–1588. McGrath, Y., Sumnall, H., Edmonds, K., McVeigh, J., Bellis, M., 2006. Review of Grey Literature on Drug Prevention Among Young People. National Institute for Health and Clinical Excellence (NICE), London. Moncada, A., Pérez, K., 2002. Reliability and validity of self-reported drug consumption in the secondary school population. Gac. Sanit. 15, 406–413 (Spanish). Morales, E., Ariza, C., Nebot, M., Pérez, A., Sánchez, F., 2008. Cannabis consumption among secondary school pupils of Barcelona (Spain): initial use, reported effects and expectancies. Gac. Sanit. 22, 321–329 (Spanish). Nebot, M., Jiménez, E., Ariza, C., Tomas, Z., 2006. Trends in the consumption of tobacco, alcohol and cannabis among adolescents of Barcelona, Spain, between 1987 and 2004. Med. Clin. (Barc) 126, 159 (Spanish). Newton, N.C., Andrews, G., Teesson, M., Vogl, L.E., 2009. Delivering prevention for alcohol and cannabis using the Internet: a cluster randomised controlled trial. Prev. Med. 48, 579–584. Pérez, A., Ariza, C., Sánchez-Martínez, F., Nebot, M., 2009. Cannabis consumption initiation among adolescents: a longitudinal study. Addict. Behav. 35, 129–134. Porath-Waller, A.J., Beasley, E., Beirness, D.J., 2010. A meta-analytic review of schoolbased prevention for cannabis use. Health Educ. Behav. 37, 709–723. Sánchez-Martínez, F., Ariza-Cardenal, C., Pérez-Giménez, A., Diéguez-Ferrer, M., López-Medina, M.J., Nebot-Adell, M., 2010. Process evaluation of a cannabis school-based prevention program “xkpts.com” in Barcelona teenagers in 2006. Adicciones 22, 217–226 (Spanish). Schafer, J., Brown, S.A., 1991. Marijuana and cocaine effect expectancies and drug use patterns. J. Consult. Clin. Psychol. 59, 558–565. Schofield, M.J., Lynagh, M., Mishra, G., 2003. Evaluation of a Health Promoting Schools program to reduce smoking in Australian secondary schools. Health Educ. Res. 18, 678–692. Sussman, S., Dent, C.W., Stacy, A.W., 2002. Project towards no drug abuse: a review of the findings and future directions. Am. J. Health Behav. 26, 354–365. Swedish Council for Information on Alcohol and other Drugs (CAN), 2009. The 2007 ESPAD Report. Substance Use Among Students in 35 European Countries. Stockholm: The European Monitoring Centre for Drugs and Drug Addiction (EMCDDA). Council of Europe, Co-operation Group to Combat Drug Abuse and Illicit Trafficking in Drugs (Pompidou Group), Stockholm. Tobler, N., Lessard, T., Marshall, D., Ochshorn, P., Roona, M., 1999. Effectiveness of school-based drug prevention programs for marijuana use. School Psychol. Int. 20, 105–137. Ventura, A., Canals, R.M., García, R., Pujol, C., Tomás, P., 1999. Family Economic Capacity Index in Barcelona City. II. Barcelona. Ajuntament de Barcelona. Gabinet tècnic de programació (Spanish). Von Sydow, K., Lieb, R., Pfister, H., Höfler, M., Wittchen, H.U., 2002. What predicts incident use of cannabis and progression to abuse and dependence? A 4-year prospective examination of risk factors in a community sample of adolescents and young adults. Drug Alcohol Depend. 68, 49–64.