Employment, Social Affairs & Inclusion

Örebro Prevention Programme

Evidence level:
 
Evidence of effectiveness:
? - 0 + ++
Transferability:
? - 0 + ++
Enduring impact:
? - 0 +

Project overview

The Örebro Prevention Programme (currently known as EFFEKT) works through parents and by targeting drinking among 13–16-year-olds (i.e. Junior high school pupils in grades 7–9). The 2.5-year programme, which was designed in Örebro County, Sweden, has been funded by the Swedish National Institute of Public Health. This has been part of its initiative for universal youth alcohol prevention programmes, which are to be implemented community-wide; target youths at ages during which drinking increases; work in both urban and rural communities; make use of existing community resources; and bring together different agencies and relevant parties.

Practice category

  • Supporting Parenting and Assisting with Childcare
  • Facilitating Positive Transitions to Adulthood

Recommendation pillars

  • Put in place mechanisms that promote children’s participation in decision making that affect their lives
  • Support the participation of all children in play, recreation, sport and cultural activities

Countries that have implemented practice

  • Sweden
  • Netherlands

Age groups

  • Teenagers (age 13 to 19)

Target groups

  • Children
  • Parents

Years in operation

  • 1995  - still operating

Scope of practice

  • local level

Type of organisation implementing practice

  • National Government

Rationale of practice

The programme design and implementation were largely based on empirical findings demonstrating a negative correlation between levels of youth alcohol drinking and (i) the strictness levels of parental attitudes against youth alcohol consumption as well as (ii) the level of youth involvement in structured, adult-led activities. Thus, the core of the programme has been based on parents receiving information (by mail and during parent meetings at the schools) encouraging them to maintain strict attitudes against youth alcohol use as well as encouraging their youth’s involvement in adult-led, organised activities. The programme was designed to be implemented in both urban and rural areas.

Mode of delivery

  • Face-to-face
  • Mail

Delivery dosage

Not available

Location of practice

  • School based

Evidence of effectiveness

Evaluation 1

The study was conducted in Sweden and its design of the first evaluation has been a quasi-experimental one, using matched controls with a pre–post, intention-to-treat design. This design collects measures both before and after the treatment, and includes all youths assigned to the treatment group, even if the youth did not get the treatment or only got a partial dose of the treatment. The treatment was such that for five semesters (i.e. 2.5 years) project workers attended parents’ meetings once per semester and provided them with information and advice. The evaluation of the effectiveness of the treatment was based on a longitudinal assessment consisting of questionnaires answered by youths in the classrooms and their parents at home 1.5 and 2.5 years after the baseline.

The measures used in the evaluation were the parental attitudes towards alcohol consumption of their children, the level of involvement of the youths in organized activities, adolescent drunkenness, and delinquency. The results have shown that the parents in the intervention group kept their strict attitudes over time, while those in the control group became more permissive. Furthermore, youth drinking and delinquency levels have increased over time to a greater extent in the control group than in the treatment one. On the other hand, the treatment seems to have no impact on the level of youth involvement in organized activities.

Koutakis, N., Stattin, H. and Kerr M., “Reducing youth alcohol drinking through a parent-targeted intervention: the Örebro Prevention Program”, Research Report, Society for the Study of Addiction, Addiction, 103, 2008, p. 1629-1637.  

Summary of Results for Evaluation 1*

Outcome

Treatment Group

Control Group

Outcomes improved (statistically significant)

Parents’ restrictive attitude

3.81

3.46

Drunkenness

0.45

0.97

Drunkenness: early starters

0.95

1.97

Delinquency

1.21

1.36

Delinquency: early starters

1.55

1.81

Outcomes with no effect

Level of youth involvement in organized activities

N.A.

N.A.

* The table presents means for intervention and control groups at post-test (i.e. 2.5 years after the programme has started) and analysis of covariance results controlling for baseline (i.e. before programme has started) in the longitudinal sample.

Evaluation 2

The second evaluation was also conducted in Sweden and designed such that 40 eligible Swedish schools (i.e. municipal schools with pupils aged 13 to 16 which were not earlier exposed to OPP and which have agreed to participate) were randomly and equally assigned to either the treatment or control groups. The treatment group has been subject to the OPP which was based on 6 20-minute long PowerPoint presentations given to parents of 13-16 years olds by trained OPP presenters. The evaluation of the treatment effectiveness was based on a longitudinal assessment consisting of questionnaires answered by youths in the classrooms and their parents at home prior to the intervention as well as 12 and 30 months after baseline.

The outcomes measured were drunkenness frequency, onset (i.e. life-time) drunkenness, alcohol consumption and parents’ attitudes towards youth drinking and alcohol served at home. The evaluation found that parents who have participated in the Orebro Prevention Programme had more restrictive attitudes toward drinking at both 12 and 30 months after baseline. However, the evaluation did not show the intervention to have any other statistically significant impacts.  This evaluation was conducted between 2007 and 2010. 

Bodin, M. C., & Strandberg, A. K. (2011). The Örebro prevention programme revisited: a cluster-randomized effectiveness trial of programme effects on youth drinking. Addiction, 106(12), 2134-2143.

Summary of Results for Evaluation 2

Outcome

T2 (i.e. 12 months after baseline)

T3 (i.e. 30 months after baseline)

Treatment Group

Control Group

Treatment Group

Control Group

Parents’ restrictive attitudes

3.86

3.74

3.78

3.56

Outcome

Odds Ratio

T2

T3

Frequent drunkenness

Completers only

0.61

0.99

Worst case scenario

0.62

0.90

Multiple imputation

0.59

1.04

Best case scenario

0.63

1.05

Life-time drunkenness

Completers only

0.94

1.07

Worst case scenario

0.82

0.98

Multiple imputation

0.95

1.07

Best case scenario

0.99

1.12

Weekly drinking

Completers only

0.90

0.93

Worst case scenario

0.75

0.85

Multiple imputation

0.89

0.99

Best case scenario

0.95

0.97

Served alcohol

Completers only

0.76

0.74

Worst case scenario

0.73

0.71

Multiple imputation

0.76

0.74

Best case scenario

0.81

0.77

Statistically significant outcomes are presented in Bold.

Evaluation 3

This evaluation consisted of a reanalysis of the previous one by Bodin and Strandberg (2011), and it uses the same data but a different statistical approach. Thus, the implementation of the intervention and the measurement process is the same as in evaluation number 2. The authors have provided the following three reasons for conducting the reanalysis: 1). the programme’s implementation largely did not reach the recommended number of sessions, 2) Bodin and Strandberg (2011) dichotomized their measures of youth drinking, which reduced the variance in the outcomes measured, and 3) the evaluation by Bodin and Strandberg (2011) did not take into account the theory behind the programme.

In order to overcome the aforementioned identified shortcomings, Özdemir and Stattin (2012) used latent growth modelling approach with the individual as the unit of analysis, using the following undichotomized variables as outcomes: life-time drunkenness, drunkenness in past month, frequency of drinking and amount of drinking on each occasion. Finally, in order to account for mediating factors the authors have also examined whether the intervention impacted parental attitudes and next whether the intervention, through this aspect, indirectly affected youth drinking.

The analysis showed the intervention to directly decrease life-time and past month drunkenness and indirectly (by increasing the levels of parental strictness) the frequency and amount drinking.

Özdemir,M. and Stattin, H. (2012) Does the Örebro Prevention Programme prevent youth drinking? Letters to the Editor. Addiction 2012; 109 (9): 1705-06.

Summary of Results for Evaluation 3

Outcome

  Est.       

  P value  

Outcomes improved (statistically significant)

Life-time drunkenness (direct)

-0.37

0.034

Past-month drunkenness (direct)

−0.352

0.054

Parental attitudes towards youth drinking (direct)

0.19

<0.01

Life-time drunkenness (indirect; through parental attitudes)

−0.057

0.016

Past-month drunkenness (indirect; through parental attitudes)

−0.072

0.017

Frequency of drinking (indirect; through parental attitudes)

−0.05

0.003

Amount of drinking (indirect; through parental attitudes)

−0.06

0.006

Outcomes with no effect

Frequency of drinking (direct)

N.A

N.A

Amount of drinking (direct)

N.A

N.A

Evaluation 4

The third evaluation was conducted between 2006 and 2008 in the Netherlands and designed as a school-level randomized controlled trial which compared the effectiveness of an amended version of OPP, a student based digital intervention and both interventions combined in reducing youth alcohol consumption. The sample used in this study consisted of 3490 students, belonging to 152 classes in 19 participating schools. Each of the schools has been randomly assigned to one of the following four groups: parent intervention (i.e. Dutch adaptation of OPP); student intervention (a renewed digital alcohol programme, aimed at helping students to develop a healthy attitude towards alcohol usage and train their refusal skills through primarily the use of e-learning); combined intervention (which constituted of both the parent and student interventions) and a control group.

The measurements were taken at baseline (prior to the start of the intervention) and 10, 22, 34 and 48 months after the baseline. The outcomes measured in the evaluation were the following: onset of heavy weekly and weekly alcohol use (with the former being the primary and the latter the secondary outcome) as well as the frequency of drinking. The data obtained were analysed using both the intent- to- treat (all students assigned to treatment) and completers only frameworks. The results obtained suggest that the combined intervention significantly delays heavy weekly drinking in the short term but not in the long term and weekly drinking and frequency of drinking in both the short and long-term run. The parent intervention (based on OPP) as well as the student intervention, when conducted separately, largely do not appear to have any significant effects on either heavy weekly drinking or weekly drinking. Therefore, in the Dutch context, only the combined Orebro prevention programme significantly impacts drinking behaviour amongst youth, and only in the short term (i.e. 10 months after baseline) as well as 34 and 48 months after baseline (i.e. it was only proven ineffective 22 months after baseline).

However, similar to the second evaluation conducted in Sweden, parents who have participated in the Orebro Prevention Programme reported more restrictive attitudes toward drinking at both 12 and 30 months after baseline
 
Summary of Results for Evaluation 4

Outcome

CI (for the parent intervention)*

T2 (i.e. 10 months after baseline)

T3 (i.e. 22 months after baseline)

Heavy weekly drinking (intention-to-treat)

0.69-2.90

0.73-1.73

Heavy weekly drinking (completers only)

0.71-3.01

0.79-1.83

Weekly drinking (intention-to-treat)

0.62-1.20

0.63-1.16

Weekly drinking (completers only)

0.55-1.34

0.43-0.95

Frequency of drinking (intention-to-treat)

-0.09-0.38

-0.44-0.27

Frequency of drinking (completers only)

-0.06-0.41

-0.45-0.33

* Reference group = control condition. Adjusted for confounders (age, level of education and sex) and cluster effect; CI; confidence interval; statistically significant outcomes are presented in Bold.

Koning I. M., Vollebergh W. A. M., Smit F., Verdurmen J. E. E., van den Eijnden R. J. J. M., ter Bogt T. F. M. et al. Preventing heavy alcohol use in adolescents (PAS): cluster randomized trial of a parent and student intervention offered separately and simultaneously. Addiction 2009; 104: 1669–78.

Evaluation 5

Given that this evaluation consists of a follow-up measurement of those students who participated in the evaluation by Koning et al. (2009) described above, the sample selection procedure is the same as described above. Namely, initially 3490 students were selected to participate in the evaluation out of which 2937 were eligible to be included in the analysis. Out of those, in 2009 (i.e. 34 months after baseline) 2533 adolescents stayed in the programme and completed the follow-up assessment resulting in a drop-out rate of 13.8%. Furthermore, the study design and the outcomes used are also the same as in the fourth evaluation (i.e. the primary and secondary outcomes are onset of heavy weekly alcohol use and weekly alcohol use, respectively).

An intention-to-treat analysis using two models (a logistic regression model and an autoregressive logistic model which included alcohol use at previous time points) was applied.

Overall, no statistically significant effects of the parent intervention (based on OPP) were reported on the incidence of heavy weekly drinking as well as on onset of weekly drinking. The combined parent-student intervention did significantly lower the onset of (heavy) weekly drinking.

Summary of Results for Evaluation 5

Outcome

Odds ratio

Outcomes with no effect

Heavy weekly drinking (model 1)

0.83

Heavy weekly drinking (model 2)

0.78

Weekly drinking (model 1)

0.86

Weekly drinking (model 2)

0.81

Model 1: Logistic multiple regression, adjusted for confounders (age, level of education, and gender) and cluster effect; Model 2: Autoregressive logistic regression analyses, adjusted for confounders (age, level of education, and gender); cluster effect; and outcome at t – 1

Koning, I.M., van den Eijnden, R.J.J.M., Verdurmen, J.E.E., Engels, R.C.M.E., Vollebergh, W.A.M. (2011). Long-Term Effects of a Parents and Student Intervention on Alcohol Use in Adolescents. American Journal of Preventive Medicine, 40, 541-547

Evaluation 6

Given that this evaluation consists of two additional follow-up measurements of those students who participated in the evaluation by Koning et al. (2009) (i.e. evaluation 4), the sample selection procedure is the same as described above.

Namely, initially 3490 students were selected to participate in the evaluation out of which 2937 were eligible to be included in the analysis. Out of those, in 2009 (i.e. 34 months after baseline; at T3) 2533 adolescents stayed in the programme and completed the follow-up assessment and out of those 1064 have participated in the follow-up 50 months after the initiation of the programme.

The study design and the outcomes used are the same as in the fourth evaluation (i.e. the primary and secondary outcomes were onset of heavy weekly and weekly alcohol use, respectively).

The analysis of the data was conducted using both the intent to-treat and completers- only frameworks; while controlling for age, gender and level of education. The results obtained have shown no overtime effect of the parent intervention (based on OPP) on the prevalence of heavy weekend drinking and weekly alcohol use. On the other hand, the parent intervention increased the parents’ strictness. However, those stricter attitudes have not translated into lower levels of heavy weekend drinking or weekly alcohol use. 

Summary of Results for Evaluation 6

Outcome

95% confidence interval

Outcomes with no effect

Prevalence of heavy weekend drinking

−0.34 to 0.36

Amount of drinking

−0.26 to 0.17

Transferability

Evaluations in the Netherlands found that the OPP does not have a statistically significant impact on youth drinking.

The programme designers, in their paper, have raised the issue of the effectiveness of the programme in countries with varying alcohol policies and different degrees of their enforcement.

Practice Materials

  • Practice materials are not available 

Cost information

  • Implementation cost information is not available

Enduring impact

It is unclear whether the intervention has had an enduring impact in Sweden as there were no follow-up measurements conducted after the completion of the programme.

In the Netherlands, Koning et al. (2013) have conducted a follow-up 50 months after baseline (i.e. 20 months after the completion of the programme) the results of the analysis have shown the intervention to be ineffective also in the long-term.

Evaluation details

Evaluation 1

For this evaluation, the selection of schools was done using data from the survey of all 9th graders (aged 16) in Örebro County (n = 3094) conducted by the Social Medicine Unit of the County Hospital and regarding alcohol use and other health-related behaviours. The authors have selected and approached intervention schools in the following three types of communities: inner city, public housing and small town. All the schools approached have agreed to participate in the programme. The pupils and parents of the selected schools were both given the possibility of not participating.  

Then, the authors went on and selected matching control schools which are as similar as possible to the intervention ones in terms of the community type they are based in, their size, the alcohol use and delinquency of the pupils but which are at the same time geographically distinct enough to minimize risks associated with cross-over effects.  

In each of the small towns two schools were chosen (one intervention and one matched control) while in each of the other community either one intervention school or one matched control school were selected per community. In total, the sample consisted of 900 pupils entering junior high school and their parents, followed longitudinally.

The participation rates for youth did not differ significantly between the intervention and control groups while they did for some of the parents’ groups.

In terms of comparability of the intervention and control groups it has been noted by the authors that parents in the former group were significantly less strict than those in the latter. There were no other significant differences in the remaining variables of interest.

As mentioned above the study has been a quasi-experimental one using matched controls with a pre–post, intention-to-treat design. The treatment was such that it covered five semesters during which project workers attended parents’ meeting once per semester. At the first of these six meetings, the workers gave 30-minute descriptions of the programme and advised parents to adopt or maintain a zero-tolerance attitude towards youth alcohol consumption and communicate clear rules to their children.

The evaluation of the effectiveness of the treatment was based on a longitudinal assessment (which included both the intervention and control group) consisting of questionnaires answered by youths in the classrooms and their parents at home 1.5 and 2.5 years after baseline. The measures used in the evaluation were the parental attitudes towards alcohol consumption of their children, the level of involvement of the youths in organized activities, adolescent drunkenness and delinquency.

In their evaluation, the authors used a multi-stage statistical analysis. First, they have examined baseline equivalence between the intervention and control groups using t-tests for independent groups. Second, in order to analyse changes in drunkenness and delinquency over time (grades 7–9) for the intervention and control groups, they used general linear model (GLM) analysis for repeated measures while controlling for youth’s gender, parental education and ethnicity as well as the community the school is based in. Additionally, the authors have also examined gender effects (through the use of interactions) and various different methods to examine the significance of the differences between the outcomes of the control and intervention groups (e.g. univariate analyses of covariance- ANCOVAs, c2 analyses or Cohen’s d measure).In terms of results, the parents in the intervention group kept their strict attitudes over time, while those in the control group became more permissive. Furthermore, while drinking increased over time for both the control and intervention groups this increase has been steeper in the former than in the latter. Similarly, an increase in delinquency for both groups has been noted over the analysed three years albeit again to a greater extent in the case of the control group. On the other hand, there is no significant difference between the intervention and control groups in the outcomes relating to the level of involvement in organized activities. Finally, the coefficients on community variables and interactions have proven statistically insignificant. It is worthwhile noting that it does not appear that the authors in their analysis have controlled for clustering of students in schools.

To control for attrition, the authors also conducted the analysis using the following two restricted samples: a sample consisting only of participants who partook in the baseline and post-test and who were matched on numbers of frequent drinkers at baseline; a sample consisting of all baseline participants for which missing post-test measures were imputed. The results obtained when using those samples are largely consistent with the ones reported for the complete sample with the exception of delinquency among early starters (the coefficient was statistically insignificant).

Evaluation 2

For the second evaluation the sample selection has been conducted in the following steps: first, the eligible schools (i.e. municipal schools in Sweden comprising the ages 13-16 with no previous exposure to the Orebro Prevention Programme) with experienced local OPP presenters within travelling distance have been identified and contacted.  Altogether, information about the study was mailed to 716 schools (in 13 counties). The authors initially sent invitations to participate to schools in six counties, and before sending the information to these schools mapping efforts of eligibility was done. However, because this procedure was very time-consuming and there was a very low response rate of schools in the first six counties, therefore the authors dropped the mapping efforts before inviting schools and subsequently checked eligibility criteria with those who responded (in the remaining seven counties).Overall, 40 schools met the eligibility criteria and agreed to participate. Those 40 schools have been stratified by cluster size as well as the proportion of graduating students (which served as a socio-economic proxy). Then, sealed envelopes containing the school’s names were assigned to groups in blocks of two (which were generated in advance by coin-tossing). This method resulted in the creation of a treatment and control groups each consisting of 20 schools. In each of the schools measurements took place in two to three classes selected by the school, overall in the former (intervention group) the sample consisted of 46 classes and 893 youth and in the latter of 41 classes and 859 pupils. Thus, in total the sample consisted of 1752 pupils who have had their initial measurements taken when in 7th grade, the follow up rates in the 8th and 9th grades were 1613 (which corresponds to 92.1% of the initial sample) and 1548 (which corresponds to 88.4% of the initial sample) respectively. The response rates recorded for parents were somewhat lower.

With regards of sample representativeness the selected schools were located in diverse municipality types with the intervention and control ones being approximately evenly distributed with each of the types. However, the grade point average recorded for the schools participating in the evaluation has been somewhat higher compared to the national average (i.e. 208.5 as opposed to 205.3).  

The study design was such that the intervention consisted of 6 20-minute long PowerPoint presentations given to parents of 13-16 years olds (during the regular termly parent-teacher meetings in grades 7 to 9) by trained OPP presenters. The presentations urged parents to adopt a strict attitude towards youth drinking as well as sign written agreements describing how youth can be prevented from drinking.

The measurement process consisted of classroom-administered youth questionnaires and questionnaires mailed to parents, which parents filled out. The administration of the questionnaires in the first round followed a pre-randomization approach (i.e. blinded assessors) while that of the latter two followed the non- blinded assessors approach. It is worthwhile noting that before the first measurement took place parents were given the option to opt out through a passive consent procedure.

With regards to outcomes measured the primary one has been drunkenness frequency while the secondary ones have been onset (i.e. life-time) drunkenness and alcohol consumption. Furthermore, parents’ attitudes towards youth drinking have been considered as an intermediary outcome.

The authors, in their paper, present the results for the two different approaches: the naive analysis which ignores the multi-level structure and the two-level logistic regression. As can be seen from table 4 below, a statistically significant programme effect (P<0.02) on frequent drunkenness has been reported only for the worst-case- scenario in the 12 month after baseline follow up (i.e. T2); for the remaining three scenarios in T2 and for all scenarios in T3 (i.e. 30 month after baseline follow up) the effects observed have been statistically insignificant. Furthermore, for both follow- ups in for all four scenarios the programme had no significant effects on life-time drunkenness or weekly drinking. Finally, parental attitudes towards youth drinking and more specifically the level of strictness has shown to be affected by the programme; namely, as can be deducted from table 4 in T2 under the worst-case-scenario and in T3 for all four scenarios the programme significantly reduced the number of alcohol served to youth at home; also, the results indicate that  parents who have been subject to OPP reported in both T2 and T3 stricter attitudes towards alcohol consumption than those belonging to the control group.   

Evaluation 3

This reanalysis of the data used by Bodin and Strandberg (2011), it uses the same study population but a different statistical approach as in the second evaluation.
Likewise, the study design (including the implementation of the intervention and the measurement process) has been the same as in evaluation number 2. The reasons, as stated by the authors Özdemir and Stattin (2012), for conducting the analysis are the following.

First, in Bodin and Strandberg’s (2011) evaluation, the programme’s implementation largely did not reach the recommended number of sessions; such incomplete implementation of the programme could have negatively impacted its effectiveness. Second, Bodin and Strandberg (2011) in their analysis dichotomized their measures of youth drinking (i.e. converted them from continues or categorical variables into nominal) which has reduced the variance in the outcomes measured. This reduction in the variance might have resulted in measures not being sensitive enough to detect differences between the treatment and control groups. Third, the evaluation by Bodin and Strandberg (2011) did not take into account the theory behind the programme. Namely, while the authors did not test any mediating mechanisms the theory states that the effectiveness of preventive interventions depends on the degree to which they manage to change the mediating mechanisms through which they are conceived to work.

In order to overcome the aforementioned identified shortcomings, Ozdemir and Stattin (2012) used latent growth modelling approach (which has been demonstrated suitable for measuring intervention effects) with the individual as the unit of analysis using the following undichotomized variables as outcomes: life-time drunkenness, drunkenness in past month, frequency of drinking and amount of drinking on each occasion. Conducting the analysis on the individual level allowed analysing to which extent the quality of the implementation affected the results while the use of undichotomized variables enabled increasing the variance between observations. Finally, in order to account for mediating factors the authors have also examined whether the intervention impacted parental attitudes and next whether the intervention, through this aspect, indirectly affected youth drinking.

The results of the first analysis (which did not take into account the mediating mechanisms) showed that the intervention had a direct significant effect on life-time drunkenness (est. = -0.37, P = 0.034) and a direct, albeit marginal, significant effect on past-month drunkenness (est. = -0.352, P = 0.054). On the other hand, though, no statistically significant effects have been reported for the frequency and amount of drinking.

In order to check whether the difference between those results and the ones reported by Bodin and Strandberg (2011), the authors have re-ran the analysis using dichotomized variables. The results obtained have shown no significant programme effects. Given those the authors concluded that the dichotomization of the outcomes in the previous evaluation is likely to have suppressed the significance of the impacts of the intervention.

The results of the two-step analysis including mediating mechanisms have shown the intervention to have statistically significant indirect effects (through parental attitudes) on each of the four outcomes measured. In more detail, the Örebro prevention programme has been shown to indirectly reduce (through its effect on parents’ attitudes toward youth drinking) life-time drunkenness (est. = -0.057, P = 0.016), past-month drunkenness (est. = -0.072, P = 0.017), frequency of drinking (est. = -0.05, P = 0.003) and amount of drinking (est. = -0.06, P = 0.006).

Evaluation 4

The sample used for the evaluation in this study consisted of 3490 students, belonging to 152 classes in 19 participating schools.

The sample selection was such that, 80 schools were randomly selected from the list containing all public secondary schools, and invited to participate in the evaluation if the three following inclusion criteria had been met: (1) the school had at least 100 first-year students; (2) less than a quarter of the students were from migrant populations; (3) the school was not offering special education.

Out of the contacted schools, the first 20 schools from various regions in the Netherlands that agreed to participate were included. Those schools have been randomly  assigned to one of the following four groups: (1) parent intervention (a Dutch adaptation of the ÖPP) (ii) student intervention (renewed digital alcohol programme which was based on the alcohol module of the Healthy School and Drugs program), (3) parent and student intervention (combined intervention) and (4) control group. After the randomization occurred, one of the schools (assigned to the control group) had to withdraw its participation due to reasons unrelated to the study.

In terms of comparability, the authors have observed statistically significant differences between the intervention groups and the control one with regards to the sex ratio, mean age and level of education which led them to control for those variables when conducting the analyses.

The study design was such that in the first group (i.e. parent intervention) the intervention was modelled on the OPP. Unlike in Sweden, however, in the Dutch evaluation the authors only focused on the outcomes related to alcohol use rather than also on engagement in organized activities; also, the meetings with the parents were integrated in the schools’ general parents meetings and their overall number was lower. In the second group (i.e. student intervention) the intervention constituted of a renewed digital alcohol programme and was based on the alcohol module of the well-established Healthy School and Drugs (HSD) Dutch prevention programme. The programme aims at helping students to develop a healthy attitude towards alcohol usage and train their refusal skills through primarily the use of e-learning. Participants belonging to the third group received both interventions (i.e. the parental and student ones) and those belonging to the fourth group (i.e. control group) received no intervention at all (with the exception of providing information about alcohol use which is a part of the Dutch curriculum).

While students were the unit of analysis, the study involved both students and parents; data from the former have been collected in the classrooms using online questionnaire and data from the latter had been collected using home mailed questionnaires. The parental and student data have been collected in three periods: before the intervention occurred (i.e. Sept/ Oct 2006) and 10, 22, 34 and 48 months after baseline (June 2007, June 2008, June 2009, September 2010 respectively).

The outcomes measured in the evaluation were the following: onset of heavy weekly and weekly alcohol use (with the former being the primary and the latter the secondary outcome) as well as the frequency of drinking.  The data obtained were analysed using both the intent- to- treat and completers only frameworks in which the outcomes of each of the experimental conditions were compared with the control one. Furthermore, in their analyses the authors also corrected for the cluster effects at class-level, given that the interventions were carried out in classes.

The results obtained suggest that the combined intervention significantly delays heavy weekly and weekly drinking in both the short (except from heavy weekly drinking at 22-month follow-up) and long term. The parent intervention (based on OPP) as well as the student intervention, when conducted separately, does not appear to have any significant effects on either heavy weekly drinking or weekly drinking. Therefore, in the Dutch context the Orebro prevention programme does not seem to significantly impact drinking behaviour amongst youth.  

Evaluation 5

Given that this evaluation consists of a follow-up measurement of those students who participated in the evaluation by Koning et al. (2009) described above, the sample selection procedure is the same as described above. Namely, initially 3490 students were selected to participate in the evaluation out of which 2937 were eligible to be included in the analysis. Out of those, in 2009 (i.e. 34 months after baseline) 2533 adolescents stayed in the programme and completed the follow-up assessment resulting in a drop-out rate of 13.8%. Likewise, the study design and the outcomes used are the same as in the previous evaluation above (i.e. the primary and secondary outcomes were onset of heavy weekly alcohol use and weekly alcohol use, respectively).

In the evaluation an intention-to-treat analysis which made use of full information likelihood to handle missing data on the dependent variables was applied. Furthermore, the analysis controlled for age, gender and level of education (as those appeared to differ amongst the groups at baseline) and corrected for the cluster effect at the class level. Finally, the analysis consisted of two models. The first one was logistic regression model which was used to obtain the ORs of weekly drinking in each of the four groups; the second one was an autoregressive logistic model (which included alcohol use at previous time points) which showed the stability or additional effect at this measurement wave compared to the previous ones included in the evaluation above.

The analysis found no statistically significant effects of either the parent or student interventions were reported on the incidence of heavy weekly drinking as well as on onset of weekly drinking.
Only the combined intervention has proven to successfully reduce the number of students who had started drinking and drinking heavily on a weekly basis (in comparison to the control group). The results reported are consistent for the two model specifications used and are also in line with the results reported in the previous evaluation conducted using the same sample.

Evaluation 6

Given that this evaluation includes an analysis of data from two additional follow-up measurements (i.e. T3 and T4) of those students who participated in the evaluation by Koning et al. (2009) described above (i.e. evaluation 4), the sample selection procedure is the same as described above. Namely, initially 3490 students were selected to participate in the evaluation out of which 2937 were eligible to be included in the analysis. Out of those, in 2009 (i.e. 34 months after baseline; at T3) 2533 adolescents stayed in the programme and completed the follow-up assessment resulting in a drop-out rate of 13.8%. Those students who have participated in the follow-up in 2009 have been asked whether they agree to participate in one more follow up (i.e. T4); a total of 1775 participants (70% of the 2009 sample) agreed to be contacted and out of those 1064 have participated in the follow-up 50 months after the initiation of the programme (which amounted to 60% of those who participated in the follow-up 34 months after initiation and 36% of those who participated at baseline).

Likewise, the study design and the outcomes used are the same as in the previous evaluation described above (i.e. the primary and secondary outcomes were onset of heavy weekly and weekly alcohol use, respectively).

The analysis of the data was conducted using both the intent to-treat and completers- only frameworks with the issue of missing data on the dependent variables being managed through the use of full information maximum likelihood (FIML).
Similarly to evaluation 5, the analysis controlled for age, gender and level of education (as those appeared to differ amongst the control and intervention groups at baseline). The regression analyses (with intervention dummies being the independent variables and the control group being the reference category) have shown the overtime effect of the parent intervention (based on OPP) on the prevalence of heavy weekend drinking to be statistically insignificant. Similarly, the parent intervention had no statistically significant impact on weekly alcohol use.

On the other hand, the parent intervention has shown to change parental attitudes as parents who were subject to this intervention reported having stricter attitudes regarding alcohol use of adolescents (est.=.19, P=.001).  However, those stricter attitudes have not translated into lower levels of heavy weekend drinking or weekly alcohol use.  Once more, the results of this evaluation, similarly to those of the two aforementioned evaluations, have shown that only the combined intervention is successful in reducing adolescent drinking.

Issues to consider

The various evaluations of the programme provide different results. While some show the intervention is effective in lowering the amount and frequency of youth drinking others show it has no statistically significant impacts on adolescent alcohol consumption. It is also important to point out that those studies which have shown positive impacts are conducted by the programme developer.

Furthermore, as outcomes are self-reported and reported by the parents there could be a bias depending on whether a student belonged to the treatment or control group.

Available resources

No available resources

Bibliography

Evaluation 1

  • Koutakis, N., Stattin, H. and Kerr M., “Reducing youth alcohol drinking through a parent-targeted intervention: the Örebro Prevention Program”, Research Report, Society for the Study of Addiction, Addiction, 103, 2008, p. 1629-1637.  

Evaluation 2

  • Bodin, M. C., & Strandberg, A. K. (2011). The Örebro prevention programme revisited: a cluster-randomized effectiveness trial of programme effects on youth drinking. Addiction, 106(12), 2134-2143.

Evaluation 3

  • Özdemir,M. and Stattin, H. (2012) DOES THE ÖREBRO PREVENTION PROGRAMME PREVENT YOUTH DRINKING? Letters to the Editor. Addiction 2012; 109 (9): 1705-06.

Evaluation 4

  • Koning I. M., Vollebergh W. A. M., Smit F., Verdurmen J. E. E., van den Eijnden R. J. J. M., ter Bogt T. F. M. et al. Preventing heavy alcohol use in adolescents (PAS): cluster randomized trial of a parent and student intervention offered separately and simultaneously. Addiction 2009; 104: 1669–78.

Evaluation 5

  • Koning, I.M., van den Eijnden, R.J.J.M., Verdurmen, J.E.E., Engels, R.C.M.E., Vollebergh, W.A.M. (2011). Long-Term Effects of a Parents and Student Intervention on Alcohol Use in Adolescents. American Journal of Preventive Medicine, 40, 541-547

Evaluation 6

  • Koning, I.M., Van den Eijnden, R.J.J.M., Verdurmen, J.E.E., Engels, R.C.M.E., Vollebergh, W.A.M. (2013). Effects of a parent and student intervention on alcohol use in adolescents four years after baseline; no evidence of catching up behavior. Addictive Behaviors, 38, 2032-2039
  • Fayers PM, Jordhoy MS and Kaasa S., Cluster-randomized trials. DOI: 10.1191/0269216302pm503xx.  Palliat Med, 2002.  
  • Gupta SK Intention-to-treat concept: A review. Perspect Clin Res 2011; 2: 109–112.

Last updated

May 2019

Find a practice (within the 'Evidence based practices'):

The practices can be also searched along the three policy pillars of the Recommendation for Investing in Children:


Share this page