+ All Categories
Home > Documents > Modifying alcohol consumption among high school students

Modifying alcohol consumption among high school students

Date post: 06-Dec-2016
Category:
Upload: sven
View: 212 times
Download: 0 times
Share this document with a friend
14
Modifying alcohol consumption among high school students An efficacy trial of an alcohol risk reduction program (PRIME for Life) Mats A ˚ . Hallgren STAD, Centre for Psychiatric Research, Karolinska Institute, Stockholm, Sweden Torbjo ¨rn Sjo ¨lund (Formerly of STAD), Aurdal, Norway Ha ˚kan Kallme ´n STAD, Centre for Psychiatric Research, Karolinska Institute, Stockholm, Sweden, and Sven Andre ´asson Swedish National Institute for Public Health, O ¨ stersund, Sweden Abstract Purpose – PRIME for Life is an alcohol risk reduction program that has been used and refined in the USA for over 20 years. A Swedish version of the program has recently been adapted for use among Swedish high-school students (age 18-19). The objective of the study is to evaluate the effects of the program on youth alcohol consumption (including high risk drinking), attitudes and knowledge about the effects of alcohol use. Design/methodology/approach – The authors conducted a randomised controlled trial involving 23 schools and 926 students. Data collection was conducted with questionnaires focusing primarily on drinking behaviour. Participants were followed up at five and 20 months to assess changes in drinking behaviour, knowledge and attitudes towards alcohol. Findings – No significant program effects on drinking behaviour were found. Knowledge about the effects of alcohol consumption on health increased after the intervention, as did negative attitudes towards alcohol, but these effects eroded over time. Originality/value – Despite being widely used in the USA and Sweden, the impact of PRIME for Life is under-reported in the literature. This is the first independent evaluation of the program focusing on high school age youth. The findings do not support the efficacy of the program as a risk reduction or behaviour change tool in a school environment. Keywords Adolescents, Alcoholic drinks, Health education Paper type Research paper Introduction Alcohol consumption and misuse by young people has been a public health concern for decades. Among adolescents, we know a great deal about the various components that The current issue and full text archive of this journal is available at www.emeraldinsight.com/0965-4283.htm This research was supported by the Swedish Social Ministry and the Swedish Council for Working Life and Social Research. HE 111,3 216 Received May 2010 Revised August 2010 Accepted October 2010 Health Education Vol. 111 No. 3, 2011 pp. 216-229 q Emerald Group Publishing Limited 0965-4283 DOI 10.1108/09654281111123493
Transcript

Modifying alcohol consumptionamong high school students

An efficacy trial of an alcohol risk reductionprogram (PRIME for Life)

Mats A. HallgrenSTAD, Centre for Psychiatric Research, Karolinska Institute,

Stockholm, Sweden

Torbjorn Sjolund(Formerly of STAD), Aurdal, Norway

Hakan KallmenSTAD, Centre for Psychiatric Research, Karolinska Institute, Stockholm,

Sweden, and

Sven AndreassonSwedish National Institute for Public Health, Ostersund, Sweden

Abstract

Purpose – PRIME for Life is an alcohol risk reduction program that has been used and refined in theUSA for over 20 years. A Swedish version of the program has recently been adapted for use amongSwedish high-school students (age 18-19). The objective of the study is to evaluate the effects of theprogram on youth alcohol consumption (including high risk drinking), attitudes and knowledge aboutthe effects of alcohol use.

Design/methodology/approach – The authors conducted a randomised controlled trial involving23 schools and 926 students. Data collection was conducted with questionnaires focusing primarily ondrinking behaviour. Participants were followed up at five and 20 months to assess changes in drinkingbehaviour, knowledge and attitudes towards alcohol.

Findings – No significant program effects on drinking behaviour were found. Knowledge about theeffects of alcohol consumption on health increased after the intervention, as did negative attitudestowards alcohol, but these effects eroded over time.

Originality/value – Despite being widely used in the USA and Sweden, the impact of PRIME forLife is under-reported in the literature. This is the first independent evaluation of the program focusingon high school age youth. The findings do not support the efficacy of the program as a risk reductionor behaviour change tool in a school environment.

Keywords Adolescents, Alcoholic drinks, Health education

Paper type Research paper

IntroductionAlcohol consumption and misuse by young people has been a public health concern fordecades. Among adolescents, we know a great deal about the various components that

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0965-4283.htm

This research was supported by the Swedish Social Ministry and the Swedish Council forWorking Life and Social Research.

HE111,3

216

Received May 2010Revised August 2010Accepted October 2010

Health EducationVol. 111 No. 3, 2011pp. 216-229q Emerald Group Publishing Limited0965-4283DOI 10.1108/09654281111123493

are necessary for effective prevention, including the risk and protective factors such asparental involvement and availability restraints (Borawski et al., 2003; Hawkins et al.,1992; Hawkins et al., 2004). For young adults aged 18-25 years, most attention hasfocused on identifiable subpopulations and/or high-risk groups, typically collegestudents. However, evidence suggests that risky drinking habits are also formed beforeuniversity or college years, when peer influences in school settings are strong(Toumbourou et al., 2009).

The present study focuses on youth in their final two years of the Swedish highschool system (18-19 years old). Although they not permitted to buy alcoholicbeverages from the Swedish retail alcohol monopoly “Systembolaget” (the age limit is20 years), they can legally be served alcohol in bars, nightclubs and restaurants. Drugprevention strategies for youth in Sweden typically focus on providing informationabout the health effects of alcohol, tobacco and other drugs (ATOD). Advances inprevention research over the past decade have indicated that while this approach mayhave a role to play in increasing public awareness and driving policy change, it hassignificant limitations when employed primarily as a behaviour change or riskreduction mechanism (Babor et al., 2010; Foxcroft et al., 2003; Skara and Sussman,2003; Tobler et al., 2000).

Despite recent indications of a reduction in total consumption levels among highschool aged youth, current drinking habits among Swedish youth are still resulting inserious harms (Centralforbundet for alkohol- och narkotikaupplysning, 2007). InSweden, Stockholm City Council annually measures the drinking habits of high schoolstudents aged 17-18 years. In 2000, 88 per cent of the students surveyed indicated thatthey had consumed alcohol at least once during the previous 12 months, while 74 percent reported having consumed six or more standard drinks in one sitting at least onceduring the previous year (Utrednings- och statistikkontoret, 2001). Two recent studiesshowed that 95 per cent of Swedish university students (aged 18-25 years) were alcoholconsumers (Sundbom, 2003) and that 34 per cent of university students under 20 yearsof age binge drank at least twice a month (Bullock, 2004). Binge drinking in particularis associated with significant short-term harm among young people, includingaccidents and assaults.

To minimise the harm caused by alcohol, it is important to assess the effectivenessof ongoing prevention efforts. The intervention in this study is a Swedish version of thePRIME for Life curriculum developed for high school students. PRIME for Life is a“risk reduction program” with education and persuasion components developed by thePrevention Research Institute (PRI) in Lexington, Kentucky. The first version of theprogram was developed in 1983 and it has been continually refined since then. Theprogram has been running in Sweden for several years, and has undergone at leasteight revisions (Prevention Research Institute, 2008). Over the years, different versionsof the program have emerged, for example for college students and driving under theinfluence (DUI) offenders, but the basic principles have remained the same. The theoryunderlying PRIME for Life is the “Lifestyle Risk Reduction Model” (Daugherty andLeukefeld, 1998) which takes into account the role played by a number of biological,social and psychological risk factors in alcohol use. It also builds on a combination ofprevailing prevention theories, for example social learning theory, the health beliefmodel and the theory of reasoned action. The model argues that alcohol and drugproblems result from the interaction between the quantity and frequency of alcohol (or

Modifyingalcohol

consumption

217

drugs) that people choose to consume and their own unique level of biological risk. Itregards alcohol misuse as a lifestyle-related health problem, with biological factors andquantity/frequency choices combining to make up the total risk for misuse (PRIME forLife, 2007). The model assumes that psychological and social triggers influence howmuch and how often an individual drinks alcohol and that modifying these identifiablerisk factors through education and persuasion will reduce the risk of future alcoholmisuse. Examples of psychological influences addressed by the model include values,attitudes, levels of stress, and personality traits such as being particularly impulsive.The program uses an interactive format and self-assessment techniques to helpparticipants recognise and articulate what it is they value most in life and to put thesein the context of their alcohol and drug choices. This process, it is argued, helpsparticipants to better understand how the alcohol choices they make can risk (orprotect) the things they value. Understanding individual triggers for alcohol use andhow biological, social and psychological factors interact to produce risk is a centraltheme within the program. PRI has a certification program for its trainers and leadersto maximise the quality of the intervention carried out.

Although the program was originally intended for use with “high risk” alcoholusers including individuals charged with DUI offences, the program has been modifiedfor different populations and settings which are likely to include individuals who drinkin risky or hazardous ways. This includes, for example, young people in the militaryand high school students. The program is underpinned by a sound theoreticalbehaviour change model and is interactive rather than didactic. Consequently, it isconceivable that the program could help reduce the risk of alcohol misuse amonghigh-school students, and not only DUI offenders.

Given that the program is widely implemented in high schools in both Sweden andthe USA, it is important to formally assess the program’s effectiveness. Althoughpopular, there is presently no empirical, peer-reviewed evidence to support theeffectiveness of the program. Hallgren et al. (2009) reported no significant programeffects on alcohol consumption or attitudes towards drinking after the program wasimplemented with young military conscripts. Moreover, in their review ofindividual-focused strategies to reduce problematic alcohol consumption by collegestudents, Larimer and Cronce (2007) called for more rigorous evaluations of PRIME forLife.

The aim of the present study was to evaluate the alcohol-preventive efficacy of thePRIME for Life curriculum among Swedish high school students, including thoseclassified as “high risk” consumers (eight or more on the Alcohol Use DisordersIdentification Test, or AUDIT). The main goal of the intervention being tested wasbehaviour change, i.e. decreased frequency of alcohol consumption, lower intake ofalcohol and decreased binge drinking. Changes in knowledge, attitudes and intentionsregarding alcohol use were also investigated, as well as perceptions of risk for alcoholproblems. In this paper, the “frequency” of alcohol use refers to the number ofoccasions on which alcohol is consumed during a given period (typically, the pastmonth), the “quantity” of alcohol used indicates how many units of alcohol wereconsumed per occasion, and “binge drinking” refers to the consumption of six or morestandards alcoholic beverages in one sitting, where a standard drink (or unit) containsaround 10 g of ethanol.

HE111,3

218

MethodThe interventionThe PRIME for Life program was implemented by trained instructors in each of the 23schools involved in the study. During a five-month period the instructors taught 24courses, with each course requiring two days, or ten hours (in two classes the coursehad to be compressed to one day). The curriculum was guided strictly by the programmanual to minimise instruction variability. After reviewing the instructors’ records ofdelivery, we were able to verify that approximately 85 per cent of the curriculum wastaught as intended, while time constraints caused the 15 per cent variation.

The high-school version used in this study emerged as a translation and culturaladaptation of the US “PRIME for Life under 21” version. This program version targetsyouth at-risk and/or subjects charged with alcohol and/or drug violations. The mostnotable differences from the adult version are a smaller exercise book and moreemphasis on youth-related issues.

DesignThe study was a group randomized trial with the PRIME for Life program used as theintervention. An age-matched control group received no intervention. There were noother potentially confounding education (or risk reduction) programs taking place atthe time of this study. Individual students responded to questionnaires administeredbefore the intervention (baseline) and at five and 20 months follow-up. Eachquestionnaire measured knowledge about alcohol, attitudes and intentions to usealcohol, different aspects of ATOD consumption (including the AUDIT questionnaire),and questions about current life circumstances. Based on an intention to treatapproach, all subjects measured at baseline were kept in the study.

MeasurementsAlcohol use was measured by the first three questions in the Alcohol Use DisorderIdentification Test (AUDIT) (Saunders et al., 1993; Saunders and Lee, 2000), Swedishversion (Bergman and Kallmen, 2002). The second question, about units consumed peroccasion, was expanded to include eight response alternatives in order to betterdiscriminate between binge and non-binge drinkers. Binge drinking was measuredwith the AUDIT question 3, “How often do you drink 6 units or more on the sameoccasion?” on a scale with five alternatives ranging from “Never” to “Daily or almostdaily”. The total AUDIT score was considered an important indicator of total alcoholuse and potential harm. Knowledge about alcohol, attitudes towards alcohol andintentions regarding alcohol use were measured with questions forming an index foreach variable. These questions were statements with five alternative responses:

(1) strongly agree;

(2) agree;

(3) indifferent;

(4) disagree; and

(5) strongly disagree.

The questions were designed to reflect common alcohol issues and issues dealt with inthe PRIME for Life program. Examples of questions were “Only people with

Modifyingalcohol

consumption

219

alcoholism in their family are at risk for developing alcoholism” (knowledge, ten items),“A party is no fun if there isn’t alcohol available” (attitudes, eight items) and “I’mplanning to cut down on my drinking” (intentions, three items). Intentions regardingalcohol use were only considered for the alcohol-consuming participants (91 per cent),but knowledge and attitudes were considered relevant even for non-drinkers. Therange of possible points was 0-4 for each index, where a higher score indicates moreknowledge or more restrictive attitudes and intentions. Perceived risk for alcoholproblems was assessed by a single question and four alternatives on a Likert-type scaleranging from “No risk . . . ” to “Very high risk for developing alcohol problems”.

Participants and procedureAll public high schools in Stockholm were available for the study. One school fordisadvantaged students was excluded. A total of 23 schools remained, eachcontributing two classes (or comparable groups), and a total of 926 students to thepresent study. Three students refused to take part in the survey. Only a few formsneeded to be discarded due to incompleteness or for providing deliberately falseresponses. The schools were stratified on location (inner city or suburban) and primaryprofile (theoretical or vocational), and then randomised to either the PRIME for Lifeintervention or the control group.

The main reason for non-participation was absence from school on the day thesurvey was administered. At baseline, only students attending school were included,but at the first follow-up, absentees were sent forms by regular mail with tworeminders. Postal questionnaires were also used for all subjects at the secondfollow-up, since by then most had left school. Table I shows the retention rate for bothconditions over time.

Statistical analysisDifferences between conditions over time were analysed with repeated-measuresANOVA and differences between conditions were analysed with t-tests. Given thatsubjects were clustered in pre-arranged groups (i.e. schools) the level of similarityamong students needed to be taken into account. We therefore calculated theintra-cluster correlation coefficient (ICC) and the corresponding variance inflationfactor (VIF) for each dependent variable and corrected the F (and t) ratios with theformula F=

pVIF. If interaction effects were found, the significance of differences was

tested with post hoc tests (Neumann- Kuhls). The statistical packages SPSS (SPSS,Chicago, IL) and Statistica (Statsoft, Tulsa, OK) were used to run these analyses.

The relevant ICCs and the corresponding variance inflation factor (VIF) werecalculated according to the general formulas for group dependency, i.e. ICC ¼ðMSbetween 2MSerrorÞ=½MSbetween þ ðm 2 1ÞMSerror� and VIF ¼ 1þ ðm 2 1 ICCÞ,

BaselineCourse

evaluationFive-monthfollow-up

20-monthfollow-up

Condition n Per cent n Per cent n Per cent n Per cent

Intervention 501 100 361 72 435 87 400 80Control 425 100 N/A 383 90 334 79Total 926 100 818 88 734 79

Table I.Retention of participantsover time, by conditionand total number

HE111,3

220

(Murray and Hannan, 1990). Due to unequal school samples the mean school size isused as m. In Table II the ICC and VIF for the primary endpoints are listed.

With an average sample of 40 students per school, 11 schools in each conditionwould give the study 80 per cent power to detect a difference of two AUDIT points ata ¼ 0:05. Similarly, 10 schools per condition would be sufficient to detect a significantdifference of 1.5 standard units of alcohol per typical drinking day. The assumedintra-class correlations in the calculations were 0.1 and 0.15, respectively.

ResultsIn total, 91 per cent of the students were alcohol consumers at the baseline survey, with2-4 drinking occasions per month the most typical frequency (reported by 53 per cent ofstudents). The majority of participants (53 per cent) indicated that they typically drankbetween three and six standard units of alcohol per drinking occasion (where onestandard unit contains about 10 g of alcohol), but 37 per cent consumed seven units ormore. Figures 1 and 2 show the frequency and quantity (units consumed per occasion)of alcohol use reported at baseline.

At baseline, the subjects were evenly distributed over the two conditions based ongeographic location (x2ð1Þ ¼ 1:99, p ¼ 0:016). A slightly skewed distribution occurredover the participants’ primary interest (theoretical/vocational) and condition(x2ð1Þ ¼ 4:86, p ¼ 0:03). Inner city students drank alcohol more often than students insuburban areas, (tð916Þ ¼ 3:03, p ¼ 0:002) but the difference between theoretical andvocational students was not significant (tð916Þ ¼ 1:16, p ¼ 0:246).

Figure 1.Frequency of alcohol use

at baseline by gender

Measure M ICC VIF

AUDIT 1 39.87 0.0858 4.335AUDIT 2 35.57 0.0944 4.263AUDIT 3 39.09 0.0869 4.049Risk for alcohol problems 39.48 0.0424 2.632Knowledge 40.09 0.0053 1.207Attitudes 40.09 0.1224 5.785Intentions 36.35 0.0501 2.771AUDIT score 34.09 0.0808 3.674

Table II.Mean number of

participants per school(M), intracluster

correlation coefficient(ICC) and variance

inflation factor (VIF) forthe primary outcomesmeasured at baseline

Modifyingalcohol

consumption

221

The mean knowledge index for PRIME For Life participants increased significantlyfrom baseline (t0) to first follow-up (t1). The effect size, as measured by Cohen’s d(Cohen, 1988) for the intervention was 0.81. There were no significant effects on thevariable “Attitudes to alcohol and intentions about alcohol use”. The perception of riskfor developing alcohol problems also increased significantly in the intervention group.The intervention effect d on perceived risk was 0.31. (See Table III for details.)

The two conditions do not differ significantly from each other on any of the fourmeasures of alcohol use from baseline to first follow-up. The overall quantity (units peroccasion) of alcohol consumed, and the total AUDIT score, is lower at five-monthfollow-up, but the differences are evenly distributed over the intervention and controlgroup. Table IV indicates the results for the alcohol use measures. The trend in

Figure 2.Units of alcohol drunk peroccasion at baseline bygender

Baseline Five monthsCondition Mean SD Mean SD Fcorr df

Knowledge (index value)Intervention 3.26 0.38 3.65 0.51 88.3 * 1,803Control 3.27 0.40 3.31 0.42

Attitudes (index value)Intervention 3.63 0.70 3.70 0.68 1.37 1,804Control 3.70 0.66 3.72 0.63

Intentions (index value)Intervention 2.11 0.76 2.23 0.72 0.953 1,702Control 2.16 0.73 2.20 0.73

Risk perception (points)Intervention 0.51 0.59 0.76 0.66 12.2 * 1,787Control 0.49 0.58 0.54 0.61

Note: *p , 0:001

Table III.Changes in knowledge,attitudes, intentions andrisk perception frombaseline to five-monthfollow-up

HE111,3

222

frequency favours the control group, while the trend for quantity of alcohol consumedfavours the intervention group.

The trends that emerged at fivemonth follow-up were maintained at the 20-monthsurvey, but were not as strong. Students in the intervention group have a significantlyhigher score on knowledge but their increase in “perception of risk” compared with thecontrol group was non-significant. The variables are described in detail in Table V.

Consistent with the five-month follow-up results, the groups did not differsignificantly on measures of alcohol consumption habits at 20 months. Although thetypical quantity of alcohol consumed declined for both conditions, the frequency ofbinge drinking declined only for the control group (although not significantly).Notably, the alcohol consumption frequency for all the students increased with 0.49SDs (tcorrð710Þ ¼ 13:16, p , 0:001) from baseline to 20-month follow up, while the

Baseline Five monthsCondition Mean SD Mean SD Fcorr df

Frequency (times/week)Intervention 0.82 0.79 0.91 0.87 0.90 1,795Control 0.77 0.82 0.78 0.82

Quantity (units/occasion)Intervention 5.95 3.38 5.35 3.28 1.79 1,684Control 5.70 3.22 5.52 3.77

Binge drinking (points)Intervention 1.47 0.93 1.40 0.96 0.10 1,698Control 1.39 0.96 1.30 0.98

AUDIT (total score)Intervention 8.05 4.54 7.15 4.11 0.09 1,647Control 7.82 4.50 6.82 4.13

Table IV.Changes in alcohol use

behaviour from baselineto five-month follow-up

Baseline 20 monthsCondition Mean SD Mean SD Fcorr df

Knowledge (index value)Intervention 3.26 0.38 3.68 0.52 41.0 * 1,716Control 3.27 0.40 3.45 0.44

Attitudes (index value)Intervention 3.63 0.70 3.81 0.65 2.24 1,715Control 3.70 0.66 3.89 0.64

Intentions (index value)Intervention 2.11 0.76 2.34 0.71 1.21 1,632Control 2.16 0.73 2.35 0.77

Risk perception (points)Intervention 0.51 0.59 0.682 0.680 2.17 1,706Control 0.49 0.58 0.534 0.631

Notes: *p , 0:001

Table V.Changes in knowledge,attitudes, intentions and

risk perception frombaseline to 20-month

follow-up

Modifyingalcohol

consumption

223

quantity declined 0.34 SDs (tcorrð621Þ ¼ 28:48, p , 0:001). Table VI shows thecondition-specific data.

Given that the PRIME for Life program was originally designed primarily to reduce“high risk” alcohol consumption, we conducted separate analyses of alcoholconsumption habits among students who obtained a total AUDIT score of 8 ormore; a clear sign of risky alcohol use. The results indicated no significant differencesbetween the intervention and control groups (Wilks’ l ¼ 0:998, Fð1 : 258Þ ¼ 0:620,p ¼ 0:432. See Table VII for mean AUDIT scores for high risk drinkers in each group.

DiscussionIndependent program evaluation is critical to ensure that limited public resources areused to maximise positive health outcomes. Despite being widely implemented in theUSA and Sweden, the present evaluation is only the second peer-reviewed study of theeffectiveness of PRIME for Life (PfL) (see Hallgren et al, 2009). The main goal of theintervention was to reduce high-risk drinking behaviour among high-school students.Although the program had some positive short term impact on student’s knowledgeabout alcohol and risks concerning heavy consumption, the primary goal of reducinghigh risk drinking was clearly not achieved. The PfL curriculum was well acceptedamong the participants. This positive reception probably contributed to theparticipants increased knowledge and awareness of their own vulnerability to

Intervention group Control groupAssessment point Mean SD Mean SD

Baseline 7.10 1.61 7.07 1.83Five-month follow-up 6.46 1.91 6.48 2.0820-month follow-up 6.38 2.09 6.25 2.02

Table VII.Means and standarddeviation scores forAUDIT C (the sum of thefirst three items onAUDIT) for high-riskalcohol consumers in theintervention and controlgroups

Baseline 20 monthsCondition Mean SD Mean SD Fcorr df

Frequency (times/week)Intervention 0.82 0.79 1.33 1.09 0.10 1,710Control 0.77 0.82 1.26 1.13

Quantity (units/occasion)Intervention 5.95 3.38 4.61 3.16 0.07 1,620Control 5.70 3.22 4.48 2.94

Binge drinking (points)Intervention 1.47 0.93 1.46 0.96 0.82 1,628Control 1.39 0.96 1.27 0.96

AUDIT (total score)Intervention 8.05 4.54 7.29 5.0 0.40 1,614Control 7.82 4.50 7.16 4.98

Table VI.Changes in alcohol usebehaviour from baselineto 20-month follow-up

HE111,3

224

alcohol problems, but did not ultimately lead to significant changes in alcoholconsumption behaviour over time.

On the whole, students in this study displayed alcohol patterns similar to adults; 10per cent were abstainers but the majority reported moderate drinking levels. Theirsomewhat high level of consumption may partly be explained by the fact that thesestudents were in a transitional period of life, during which their alcohol habitsprobably had not stabilised. Peer pressures, increased social activities and access toalcohol in bars and nightclubs may also have inflated consumption levels during thisperiod. Some participants reported problems associated with their alcohol use and, inthis respect, they are comparable to people with an established pattern of misuse.

Interestingly, this study has shown that even though the frequency of alcohol useincreased, the quantity of use (units per occasion) and frequency of binge drinking didnot. The data does not show whether the increase in frequency is attributable to thestudents who turned 20 years of age (the age at which alcohol may be legallypurchased through the government-run “Systembolaget” in Sweden). However, it isplausible that the overall availability did increase, since some of them had attained thelegal age for buying alcohol by 20-month follow up.

Some students were absent on the day of the baseline questionnaire and hence werenot reached by this study at all. The proportion was approximately three or fourstudents per school, i.e. less than 10 per cent of the target group. Some students wereabsent due to illness and some had abandoned their studies, but were still in the schoolrecords, i.e. the proportion of “true” absenteeism was even less.

The reliability of confidential questionnaires has been described thoroughly (Denniset al., 2004; O’Malley et al., 2000; Smith et al., 1995) and supports studies of this kind.Although sometimes questioned (Strunin, 2001), the reliability of measuring alcoholhabits using closed questions in self-report forms has been shown to be as reliable asusing open-ended questions (Brener et al., 1995; Lintonen et al., 2004).

One potential weakness of the study is the questionable parametric properties of theknowledge scales used. Choosing measurements to assess change in knowledge, and toa lesser degree attitudes and intentions, was a challenge in the sense that the questionsneeded to be both generally applicable and at the same time sensitive to the PRIME forLife curriculum. The solution was a product of our own, developed by an expert alcoholand drug research team, with the strengths and weaknesses of any new measurement.Pilot testing assured high face validity and experts’ opinions were used to validate thecontent. The AUDIT, on the other hand, is one of the most widely used questionnairesto assess alcohol consumption among adults and youth, and is currently used tomeasure alcohol habits in Sweden (Kallmen et al., 2007). The psychometric propertiesof the AUDIT have been reported elsewhere (Bergman and Kallmen, 2002). Thequestionnaire was developed with the ambition that it would be useful in countrieswith differing alcohol habits and cultures.

In the present study, we made strong efforts to separate the evaluation from theintervention. Still, to investigate is to some extent also to intervene and the purpose ofthe evaluation was evident, even if not outspoken, to students in the interventiongroup. By randomizing schools, we reduced the risk of contamination betweenconditions, although subjects, as well as teachers from different schools, occasionallyinteract in non-school settings.

Modifyingalcohol

consumption

225

Stockholm is the largest metropolitan area in Sweden and the results are not directlygeneralisable to the rest of Sweden, although the similarities between regions are usuallyregarded as larger than the differences. The comprehensiveness of the study, includingall public gymnasium schools in the community, assures high representation of differentsocio-economic groups and cultures as well as school-specific customs and values.

Concluding remarksAt the global level, alcohol is the third highest contributor to disability adjusted lifeyears (World Health Organization, 2007). In Sweden and many other Europeancountries, overall consumption levels, including high risk drinking and relateddamage, are rising (Anderson and Baumberg, 2006). These concerning trends contrastwith the high popularity but generally low impact of education and persuasioninterventions such as PRIME for Life. Given the poor track record of most educationinterventions, relying heavily on them is unlikely to reduce the considerable burden ofalcohol-related harms among young people or adults.

The immense challenges associated with changing alcohol use behaviour are wellknown and at least partly understood (Bullock, 2004; Foxcroft et al., 2003). Whileeducation strategies such as PRIME for Life may influence short-term attitudes andknowledge, it must be emphasised that they are rarely associated with long-termchanges in alcohol and drug use behaviour (Babor et al., 2010). In the case of school-basededucation and persuasion programs, a review by Foxcroft et al. (2002) found that afterscreening 600 papers, reports and dissertations, only two demonstrated positive findings– a social marketing based media intervention and a study pertaining to theStrengthening Families Program. If behaviour change and risk reduction is the goal, thenstrategies that affect the price, availability andmarketing of alcohol to young people (andadults) have far superior efficacy, and this is now well established (Babor et al., 2010).Strategies that are both effective and cost-effective should, in the first instance, be thepriority. That said, it is important not to dismiss the positive intention of programs suchas PRIME for Life, and to recognise their value as one way to inform the general publicand high-risk populations about the consequences of alcohol consumption and misuse.Strategies that aim to inform or persuade have the potential to initiate a critical shift incommunity attitudes that can trigger the impetus necessary for health policy reform.Such programs are also a public statement about community values and social norms.

In our view, education and persuasion strategies potentially have a place in arationally devised prevention initiative driven by availability constraints and changesto high-risk drinking environments (Graham and Homel, 2008) – but their role shouldbe reprioritised and our expectations clarified. As Norman Giesbrecht (2007) hassuggested, a reframing of these interventions might include:

. providing information to policy makers about the global burden from alcoholand related costs, and the potential impacts of specific interventions in reducingthis burden;

. informing the public of the advantages of a population based and environmentalchange approach to reducing alcohol problems and what works in reducingalcohol related harm among young people and adults; and

. telling the public about damage from alcohol, particularly damage that is notwidely known, such as alcohol-related cancers.

HE111,3

226

In Sweden and elsewhere there is wide public support for a range of alcohol-controlmeasures. This support is an important resource to be mobilised in a reframedapproach to alcohol education. It also appears likely that some interventions will not beeliminated completely, despite their apparent ineffectiveness. Therefore, a priorityshould be to explore options for increasing their impact and ensuring independentassessment of their effects. Research exploring the assumption that educationincreases the overall effectiveness of multi-component interventions also needs to beexamined closely. Where high-school students are concerned, the school educationcurriculum pertaining to alcohol could be reshaped significantly in many countries.This could involve focussing on raising awareness of alcohol policy options, teachingstudents about the key players in alcohol policy debates, and indicating the role thatstudents and parents can play in reshaping alcohol policies (Giesbrecht, 2007). Thetransition to more effective interventions is likely to be a slow process. To facilitatethis, we urge funding agencies to phase out interventions that are not effective and togive priority to those that really work.

References

Anderson, P. and Baumberg, B. (2006), Alcohol in Europe, Institute for Alcohol Studies, London.

Babor, T., Holder, H., Caetano, R., Homel, R., Casswell, S., Livingston, M., Edwards, G.,Osterberg, E. and Giesbrecht, N. (2010), Alcohol: No Ordinary Commodity – Research andPublic Policy, Oxford University Press, Oxford.

Bergman, H. and Kallmen, H. (2002), “Alcohol use among Swedes and a psychometric evaluationof the alcohol use disorders identification test”, Alcohol and Alcoholism, Vol. 37 No. 3,pp. 245-51.

Borawski, E.A., Levers-Landis, C.E., Lovegreen, L.D. and Trapl, E.S. (2003), “Parentalmonitoring, negotiated unsupervised time, and parental trust: the role of perceivedparenting practices in adolescent health risk behaviors”, Journal of Adolescent Health,Vol. 33 No. 2, pp. 60-70.

Brener, N.D., Collins, J.L., Kann, L., Warren, C.W. and Williams, B.I. (1995), “Reliability of theYouth Risk Behavior Survey Questionnaire”, American Journal of Epidemiology, Vol. 141No. 6, pp. 575-80.

Bullock, S. (2004), “Alcohol, drugs and student lifestyle! A study of the attitudes, beliefs and useof alcohol and drugs among Swedish university students”, Research Report No. 21,SoRAD, Stockholm.

Centralforbundet for alkohol- och narkotikaupplysning (2007), Rapport No. 107,Centralforbundet for alkohol- och narkotikaupplysning (Central Agency for Alcohol andDrug Information), Stockholm.

Cohen, J. (1988), Statistical Power Analysis for the Behavioral Sciences, 2nd ed., LawrenceErlbaum Associates, Hillsdale, NJ.

Daugherty, R.P. and Leukefeld, C. (1998), Reducing the Risks for Substance Abuse – A LifespanApproach, Plenum Press, New York, NY.

Dennis, M.L., Funk, R., Godley, S.H., Godley, M.D. and Waldron, H. (2004), “Cross-validation ofthe alcohol and cannabis use measures in the Global Appraisal of Individual Needs (GAIN)and Timeline Followback (TLFB; Form 90) among adolescents in substance abusetreatment”, Addiction, Vol. 99, Supplement 2, pp. 120-8.

Foxcroft, D.R., Ireland, D., Lowe, G. and Breen, R. (2002), “Primary prevention for alcohol misusein young people: a systematic review”, Cochrane Database System Review, No. 3.

Modifyingalcohol

consumption

227

Foxcroft, D.R., Ireland, D., Lister-Sharp, D.J., Lowe, G. and Breen, R. (2003), “Longer-termprimary prevention for alcohol misuse in young people: a systematic review”, Addiction,Vol. 98 No. 4, pp. 397-411.

Giesbrecht, N. (2007), “Reducing alcohol-related damage in populations: rethinking the roles ofeducation and persuasion interventions”, Addiction, Vol. 102 No. 9, pp. 1345-9.

Graham, K. and Homel, R. (2008), Raising the Bar: Preventing Aggression in and around Bars,Pubs and Clubs, Willan, Portland, OR.

Hallgren, M.A., Kallmen, H., Leifman, H., Sjolund, T. and Andreasson, S. (2009), “Evaluation ofan alcohol risk reduction program (PRIME for Life) in young Swedish military conscripts”,Health Education, Vol. 109 No. 2, pp. 155-68.

Hawkins, J.D., Catalano, R.F. and Miller, J.Y. (1992), “Risk and protective factors for alcohol andother drug problems in adolescence and early adulthood: implications for substance abuseprevention”, Psychological Bulletin, Vol. 112 No. 1, pp. 64-105.

Hawkins, J.D., Van Horn, M.L. and Arthur, M.W. (2004), “Community variation in risk andprotective factors and substance use outcomes”, Prevention Science, Vol. 5 No. 4, pp. 213-20.

Kallmen, H., Wenneberg, P., Berman, A. and Bergman, H. (2007), “Alcohol habits in Swedenduring 1997-2005 measured with the AUDIT”, Nordic Journal of Psychiatry, Vol. 61 No. 6,pp. 466-70.

Larimer, M.E. and Cronce, J.M. (2007), “Identification, prevention, and treatment revisited:individual-focused college drinking prevention strategies 1999-2006”, AddictiveBehaviours, Vol. 32 No. 11, pp. 2439-68.

Lintonen, T., Ahlstrom, S. and Metso, L. (2004), “The reliability of self-reported drinking inadolescence”, Alcohol and Alcoholism, Vol. 39 No. 4, pp. 362-8.

Murray, D.M. and Hannan, P.J. (1990), “Planning for the appropriate analysis in school-baseddrug-use prevention studies”, Journal of Consulting and Clinical Psychology, Vol. 58 No. 4,pp. 458-68.

O’Malley, P.M., Johnston, L.D., Bachman, J.G. and Schulenberg, J. (2000), “A comparison ofconfidential versus anonymous survey procedures: effects on reporting of drug use andrelated attitudes and beliefs in a national study of students”, Journal of Drug Issues, Vol. 30No. 1, pp. 35-54.

Prevention Research Institute (2008), Prime Times, January, available at: www.primeforlife.org/assets/pdf/newsletters/winter%2008.pdf (accessed 13 July 2009).

PRIME for Life (2007), Instructors’ Workbook, PRI Nordi, Lidingo.

Saunders, J.B. and Lee, N.K. (2000), “Hazardous alcohol use: its delineation as a sub-thresholddisorder, and approaches to its diagnosis and management”, Comprehensive Psychiatry,Vol. 41 No. 2, Supplement 1, pp. 95-103.

Saunders, J.B., Aasland, O.G., Babor, T.F., de la Fuente, J.R. and Grant, M. (1993), “Developmentof the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project onearly detection of persons with harmful alcohol consumption – II”, Addiction, Vol. 88 No. 6,pp. 791-804.

Skara, S. and Sussman, S. (2003), “A review of 25 long-term adolescent tobacco and other druguse prevention program evaluations”, Prevention Medicine, Vol. 37 No. 5, pp. 451-74.

Smith, G.T., McCarthy, D.M. and Goldman, M.S. (1995), “Self-reported drinking andalcohol-related problems among early adolescents: dimensionality and validity over 24months”, Journal of Studies in Alcohol, Vol. 56 No. 4, pp. 383-94.

Strunin, L. (2001), “Assessing alcohol consumption: developments from qualitative researchmethods”, Social Science in Medicine, Vol. 53 No. 2, pp. 215-26.

HE111,3

228

Sundbom, L. (2003), Studenters alkoholbruk, SAMU, Uppsala (in Swedish).

Tobler, N.S., Roona, M.R., Ochshorn, P., Marshall, D.G., Streke, A.V. and Stackpole, K.M. (2000),“School-based adolescent drug prevention programs: 1998 meta-analysis”, Journal ofPrimary Prevention, Vol. 20 No. 4, pp. 275-336.

Toumbourou, J.W., Hemphill, S.A., McMorris, B.J., Catalano, R.F. and Patton, G.C. (2009),“Alcohol use and related harms in school students in the USA and Australia”, HealthPromotion International, Vol. 24 No. 4, pp. 373-82.

Utrednings- och statistikkontoret (2002), Drogvaneundersokning i Stockholms skolor ar 2002,Utrednings- och statistikkontoret, Stockholm (in Swedish).

World Health Organization (2007), WHO Expert Committee on Problems Related to Alcohol,2007, WHO Technical Report Series 944, Second Report, Vol. 944, World HealthOrganization, Geneva.

Corresponding authorMats A Hallgren can be contacted at: [email protected]

Modifyingalcohol

consumption

229

To purchase reprints of this article please e-mail: [email protected] visit our web site for further details: www.emeraldinsight.com/reprints


Recommended