Talk About Alcohol: impact of a school-based alcohol intervention on early adolescents

Abstract

Despite a downward trend in alcohol consumption among 11–15-year-olds in recent years (Fuller, E, ed. 2013. Smoking, Drinking and Drug Use Among Young People in England in 2012. London: Health and Social Care Information Centre. http://www.hscic.gov.uk/catalogue/PUB11334), the proportion of young people seeking support from alcohol-related specialist services, and the societal and monetary costs associated with alcohol consumption, need to be addressed. Education can play an important role in this. The evaluation of the Alcohol Education Trust's Talk About Alcohol school-based intervention was conducted across England between November 2011 and October 2013 by independent evaluators. The aim was to compare the alcohol-related knowledge, awareness, and behaviour of students aged 12–14 in an intervention group with a statistically matched comparison group. Three identical surveys were carried out with approximately 4000 students to explore change over time over a 16–18-month timeframe. Multilevel modelling looked at changes in outcomes over the three time points and control for measured differences between intervention and comparison groups. There was evidence of a statistically significant delay in the age at which teenagers start to drink. There was also a significant association between the intervention and knowledge of alcohol and its effects. Although levels of frequency of drinking and binge drinking were lower than in the comparison schools, there were no statistically significant differences. Students from both groups identified personal, social, and health education lessons as a preferred source of information about alcohol and its effects. The positive impact on alcohol knowledge and the delayed onset of drinking show that the materials may support England's policy priorities around alcohol.

Keywords::

• alcohol education
• early adolescents
• secondary school intervention
• frequency of drinking
• binge drinking
• PSHE
• knowledge about alcohol
• onset of drinking
• longitudinal research
• multilevel modelling

Notes

^1. Public Health England is the national agency for protecting and improving the nation's health and well-being and tackling health inequalities.

^2. The Good Behaviour Game does not have an alcohol-specific focus. However, there is evidence that by keeping children engaged and improving behaviour in the classroom, it can significantly reduce later anti-social behaviour including the probability of developing any symptoms of alcohol abuse (James 2011).

^3. The model included all students with a valid response to the question ‘have you ever had a whole alcoholic drink’; a total of 12,409 observations. Background characteristics for cases that had missing values were imputed to have the overall sample average. A logistic model containing three levels: time, student, and school, was used. Variables were centred (continuous variables each had a mean of zero) to enable ready interpretation of the intercept term. Backwards selection was used to remove background characteristics from the model that were not statistically significant.

^4. In order to do this, the individual ethnicity variables were replaced with a ‘non-white’ dummy variable (the default is white). Backwards selection was re-run and extra interactions included in the model to measure the differential effects.

^5. The model included all students with a valid score on the knowledge variable; a total of 12,263 observations. Background characteristics for cases that had missing values were imputed to have the overall sample average. The model contained three levels: time, student, and school. Explanatory variables were centred (continuous variables each had a mean of zero) to enable ready interpretation of the intercept term. Backwards selection was used to remove background characteristics from the model that were not statistically significant.